1. Introduction
Smart cities have rapidly evolved from being a futuristic idea to a fundamental approach for tackling urban development and urbanisation challenges [1,2,3]. A smart city is one where traditional networks and services are enhanced using digital and telecommunication technologies to improve the efficiency and quality of life for its inhabitants while meeting the economic, social, and environmental needs of the present and future generations [4]. The development of smart cities is heavily influenced by key technologies, such as the Internet of Things (IoT) [5], Artificial Intelligence (AI) [2,6], cloud and edge computing [7,8], and big data analytics [9]. These technologies enable cities to monitor and manage urban services better, anticipate future demands, and respond to potential problems that may arise.
However, as urban populations grow, traditional services are becoming increasingly inadequate in meeting the ever-changing needs of urban environments. This necessitates the need for adaptive urban solutions that leverage real-time data and AI-driven analytics to enhance urban services by continuously monitoring and adjusting operations [10,11] such as traffic flow [12], energy usage [13], waste management [14], and emergency responses [15]. Such solutions can dynamically adjust to new temporal and spatial environmental urban conditions, enabling efficient resource utilisation, reducing operational costs, and enhancing service delivery [11]. In a smart city, data from the entire urban landscape is gathered and analysed in real time through IoT sensors and devices [16]. Analysing these data in real-time facilitates the detection of problems and opportunities, allowing for swift responses to the dynamic urban environment conditions. Consequently, this offers tailored solutions through data-driven decision-making processes with the ability to learn and adapt to new contexts.
In addition to real-time data analytics, the metaverse has emerged as the future of the Internet, where digital and physical worlds converge, offering a new layer of technological integration for smart cities with the potential to enhance urban sustainability. The term metaverse, derived from the prefix “meta”, meaning “beyond”, and the suffix “universe”, was introduced by Neil Stephenson in his 1992 science fiction novel Snow Crash [17]. It refers to a digitally simulated environment intricately connected to the physical world, where individuals can communicate, interact, and explore using digital avatars and immersive technologies such as virtual reality (VR) devices like head-mounted displays (HMDs), VR headsets, and smart glasses [18,19]. Rather than being limited to a single platform or technology, the metaverse is composed of a wide range of interconnected virtual environments [20]. Over time, the metaverse has evolved into an interactive space that integrates extended reality (XR) devices—including VR, mixed reality (MR), and augmented reality (AR) [21], alongside wireless communication networks, making it ubiquitous [22]. As a result, users are able to engage in virtual experiences that blur the boundaries between virtual and physical realities, offering a heightened sense of presence and immersion [19].
In recent years, the metaverse has emerged as a transformative virtual interconnected space that integrates AI with a comprehensive understanding of spatiotemporal and semantic information [22,23]. This integration enables seamless navigation, interaction, and immersive experiences in AR/VR environments across various platforms and devices [24]. The metaverse facilitates efficient communication among individuals and devices, allowing for extensive, real-time data transmission and AI-driven analysis crucial for natural language processing (NLP), information retrieval, and knowledge representation [25]. AI techniques are used to understand the meaning and the context of the real-time data being exchanged [26,27]. This is particularly useful in adaptive urban systems as processing the spatial and temporal aspects of the data based on IoT devices such as air quality sensors, traffic cameras, and energy meters can help identify patterns such as poor air quality, congested areas, and high energy consumption, respectively [22].
In addition, AI plays a crucial role in enabling key technologies, such as digital twins (DTs), 5G/6G, and cloud and edge computing, to realise the metaverse in a smart city. For instance, AI is used to analyse the real-time data collected from IoT devices deployed in the city and structured by the DTs. For each physical entity in the city, such as buildings, roads, or even the entire city, a digital representation of it is created in the metaverse, mirroring its physical counterparts using the data from the IoT devices. Any changes in the physical world are also reflected in the digital world, allowing for fast and effective city management within the metaverse. For instance, smart parking systems can leverage the metaverse to create virtual representations of parking lots, providing real-time monitoring and the efficient management of parking resources [28]. Similarly, navigation systems can offer personalised and real-time guidance through visual overlays and voice instructions, enhancing the overall transportation experience [29]. Additionally, AI contributes to network optimisation by managing resources, optimising bandwidth allocation, reducing latency to ensure seamless connectivity, and intelligently distributing computational tasks between cloud and edge nodes. Machine learning (ML) algorithms can predict future patterns and make data-driven decisions, allowing city management to manage services across the city proactively based on predicted demand.
However, while AI offers significant benefits, it also presents challenges that must be addressed to ensure the seamless coexistence of humans and machines, both in physical smart cities and their digital counterparts in the metaverse. For instance, to develop animated 3D building models, state-of-the-art computer vision can be utilised. However, utilising the 3D content for the metaverse is challenging due to the scalability of the current infrastructure. Therefore, it is essential to develop tools and infrastructure that allow developers to create more scalable 3D/AR/VR experiences across various platforms and purposes [30]. In addition, as the metaverse is massive and creates interpretability challenges, developing explainable smart city applications is essential [31]. The metaverse will rely on explainable AI to ensure transparency and trustworthy, immersive experiences across smart city applications. Addressing privacy and security within the metaverse to protect user data is also essential. Furthermore, establishing performance measurement and standardisation helps create consistent evaluation metrics, benchmarks, and best practices, facilitating the interoperability and reliability of metaverse applications.
Despite the potential of the metaverse in transforming smart city applications, there is a lack of comprehensive reviews addressing the advancement and role of AI techniques and AI-enabled technologies in the context of smart cities. This paper aims to fill this gap by investigating advanced AI techniques and key enabling technologies in the metaverse that contribute to adaptive and sustainable smart city services. This review focuses on answering the following research questions:
(1). What foundational AI techniques, advancements, and technologies enable the metaverse, and how do they enhance its ability to gather, interpret, and process data for improved urban decision making and personalised user experiences in a smart city? How are they evaluated?
(2). What are some AI applications in the metaverse for smart cities, and how are they utilised across various smart city domains—such as mobility, energy, health, environment, governance, and the economy—to contribute to adaptive and sustainable urban services? What are the notable metaverse smart city use cases in the real world, and what are their key findings and contributions?
(3). What are the major challenges and future research directions in deploying metaverse technologies for smart cities, particularly concerning scalability, interoperability, privacy, security, ethical considerations, and performance measurements?
1.1. Related Surveys
In this subsection, we examine the latest surveys on various technologies related to smart cities. Numerous studies have recently been conducted to explore the potential of smart city technologies, their implementation, applications, and future research directions [6,19,29,32,33,34,35,36,37,38,39,40,41].
For instance, Wang et al. [32] present a comprehensive framework for integrating Artificial General Intelligence (AGI) and parallel intelligence into metaverse-based smart cities, emphasising human ethics, social responsibility, and ecological sustainability. Similarly, the authors in [33] examine the ethical implications of the metaverse as a form of data-driven smart cities, focusing on privacy, surveillance capitalism, data surveillance, geo-surveillance, human health and wellness, and collective cognitive echo chambers. The researchers in [35] also provide an extensive overview of the potential and already implemented metaverse applications in smart cities, discussing various benefits and challenges associated with these applications. Similarly, ref. [29] highlights the metaverse’s potential to enhance environmental, economic, and social sustainability in smart cities, focusing on integrating digital twins and AI technologies. Bibri et al. [34] provide a comprehensive overview of the potential applications, opportunities, and challenges associated with deploying XR technologies in IoT applications within the broader framework of the Internet of City Things (IoCT), emphasising the synergy between XR and Artificial Intelligence of Things (AIoT) technologies. Furthermore, the integration of metaverse technologies into intelligent transportation systems is studied, highlighting their potential to enhance transportation safety, reliability, and efficiency through secure communication, virtual simulations, and real-time analytics [19]. The role of AI, including machine learning algorithms and deep learning architectures, in the foundation and development of the metaverse is explored, providing a comprehensive investigation of AI-based methods concerning several technical aspects and AI-aided applications, such as healthcare, manufacturing, smart cities, and gaming, in virtual worlds in [40]. Similarly, Chen et al. [41] provide a detailed review of smart cities based on metaverse technologies. Alahi et al. [39] discuss the role of IoT and AI in developing smart cities, focusing on recent advancements, potential applications, and future trends, emphasising integrating these technologies to enhance urban sustainability and productivity. Yaqoob et al. [38] reviewed the enabling technologies, opportunities, challenges, and future directions of the metaverse for smart cities, highlighting its potential to enhance infrastructure, services, and sustainability. Arora et al. [6] examine the role of multi-agent systems in smart city applications, providing a detailed description of AI paradigms, critical application areas, and future research directions, while the researchers in [36] study the integration of advanced technologies for sustainable smart cities, focusing on the role of IoT, AI, blockchain, and other technologies in enhancing various smart city domains and addressing sustainability challenges.
Unlike previous works, our study reviews current AI advancements such as generative Artificial Intelligence (GAI), including large language models (LLMs). Moreover, it focuses on sustainability, addressing critical applications like mobility, environmental and energy management, and optimisation. An extensive range of technologies, including big data analytics, NLP, computer vision (CV), IoT, digital twins, blockchain, 5G/6G, and cloud/edge computing, is also covered. Additionally, our study briefly reviews the performance evaluation criteria for metaverse applications for smart cities and discusses some metaverse implementations in real-world smart cities. A summary of these studies compared to our study is presented in Table 1.
1.2. Method
The methodology adopted in this review followed the widely recognised PRISMA methodology [42], commonly used for systematic reviews. A comprehensive search was conducted across multiple academic databases and other reputable sources, such as websites and industrial reports, as illustrated in Figure 1. Specific keywords were employed, including “AI for Metaverse”, “GAI OR LLM for Metaverse”, “digital twins”, “blockchain for Metaverse”, “explainable AI”, “IoT enabled Metaverse”, “6G powered for the Metaverse”, “edge and cloud computing for the Metaverse”, and “smart city” based on their frequent use in the recent literature and their alignment with the study objectives. These keywords were then combined with terms relevant to urban technology and sustainability, such as “Adaptive urban Systems”, “Smart Infrastructure”, “Urban Planning”, and “Sustainable Cities”. For example, the search string ‘(“GAI for Metaverse” OR “digital twins”) AND (“smart city” OR “adaptive urban systems” OR “sustainability”)’ ensured a focus on relevant intersections while capturing diverse sources. The selected articles comprised early-access materials and published peer-reviewed research from technical conferences and journals. For screening, we focused on the core elements of each paper relevant to our research topic by screening titles, abstracts, conclusions, and introductions.
The review applied strict inclusion criteria, considering only peer-reviewed papers published in English within the last ten years (Additional studies are referenced to provide foundational concepts and methodological insights and might fall outside the specified time range). Emphasis was placed on the studies that examined the integration of the metaverse, AI, and smart city, covering applications, theoretical frameworks, technologies, case studies, challenges, and future directions. An initial search yielded 1852 records from academic databases as depicted in Figure 1. Additional records (n = 90) were identified from other sources, bringing the total to 1942 records. After removing duplicate entries, 981 records remained for screening. Out of these, 57% were deemed irrelevant or limited in full content access, leading to the exclusion of 561 articles. This left 420 articles from reputable databases which were assessed for eligibility. Applying our refined research questions further narrowed down the selection to 284 papers for analysis as shown in Table 2, with 90% of the papers undergoing qualitative analysis.
1.3. Contributions
It is clear from the above Section 1.1 that some of the existing surveys have focused on very narrow perspectives of AI and relevant technologies, while others have investigated the role of the metaverse in terms of societal, economic, and digital value, as well as covering limited applications in the context of sustainable smart cities in the metaverse. This review primarily focuses on the role of AI in enabling metaverse technologies for smart city applications. Specifically, we analyse how AI techniques like ML, DL, GAI, and NLP integrate with the metaverse in smart cities. In addition, we explore AI’s contributions to technologies like digital twins, big data analytics, and edge computing, demonstrating their role in delivering adaptive and sustainable solutions for smart cities. Practical applications and real-world use cases are also included.
Our contributions are as follows:
(1). We examine the key AI techniques and technologies enabling the metaverse, including GAI/LLM, NLP, ML, CV, and big data analytics. We analyse how each technology enhances the metaverse for improved urban decision making and personalised user experiences. Additionally, we explore AI’s role in enabling technologies such as digital twins, IoT, edge/cloud computing, and 5G/6G communication.
(2). We outline various applications within the metaverse that contribute to adaptive and sustainable smart city services across smart environments, mobility, energy, health, governance, and the economy. We demonstrate how these applications leverage AI for context-aware, data-driven urban planning and management decision making through examples. We also present notable real-world use cases of the metaverse in real-world smart cities.
(3). We identify and discuss major challenges in deploying metaverse applications for smart cities, such as scalability, interoperability, privacy and security, ethical considerations, and performance measurement. We also explore future research directions focused on developing robust frameworks, standardising evaluation metrics, and establishing ethical AI practices to ensure an inclusive, interoperable, and sustainable metaverse for smart city applications. We propose a generic conceptual model based on the Technology Acceptance Model (TAM) for the metaverse in smart city applications.
The rest of the article is organised as follows: Section 2 includes the fundamentals and background of smart cities and the metaverse. Section 3 presents the state-of-the-art AI and its role in realising the metaverse and enabling key supporting technologies. Section 4 highlights the integrated role of AI and key technologies in realising the metaverse for sustainable smart cities, along with potential applications and use cases. Section 5 discusses the implications, challenges, and future directions. Section 6 concludes the paper.
2. Fundamentals and Background
Smart cities are in a constant state of evolution, applying cutting-edge information and communication technologies (ICTs) to enhance the management of urban services while fostering reliability and sustainability [43]. Understanding that a city functions as a multifaceted system with various stakeholders operating in uncertain environments [44], smart cities should be regarded as complex adaptive entities that self-organise and evolve, continuously acquiring new service-oriented knowledge [45]. This paradigm shift underscores the importance of adaptive urban systems, which harness advanced technologies to create more efficient, sustainable, and liveable urban environments [4,11], contributing to achieving the UN Sustainable Development Goals (
In adaptive urban systems, urban services are characterised by their ability to respond and adjust to continuous changes and disruptions within a smart city [11,46]. They employ data and connectivity to execute urban functions, addressing major challenges faced by modern cities [47]. For example, intelligent transportation systems optimise traffic flow and public transit schedules [12], lowering travel times and pollution levels. Similarly, smart energy [48] and waste management systems [49] help to ensure sustainability by efficiently distributing energy based on real-time demand and using sensor-equipped bins with data-driven pickup routes. These services improve citizens’ convenience, safety, and accessibility, resulting in an improved quality of urban life.
To ensure that smart cities are adaptable, scalable, and capable of evolving in response to real-time data, the architecture must encapsulate essential elements that enable urban services to dynamically adjust to the changing needs of the urban environment and its citizens. Based on the literature, we identify a reference architecture that integrates advanced technologies and data-driven decision-making processes, structured into four primary layers [50]:
(1). Sensing Layer: This foundational layer collects real-time data from physical devices such as sensors and IoT devices distributed throughout the city [51]. The ability of IoT devices to receive updates without physical intervention ensures the sensor network remains agile and responsive to urban developments [52].
(2). Transmission Layer: Facilitates communication between the sensing layer and upper layers by transmitting the collected data through various communication technologies.
(3). Data Management Layer: Processes and stores the transmitted data, preparing it for analysis and meaningful interpretations that inform smart city services. Adaptive urban systems may also leverage spatial and temporal data for context-aware analytics [53].
(4). Application Layer: Utilises the processed data to deliver smart city services such as transportation, energy, and waste management. Security modules are integrated into each layer to protect sensitive data, addressing key concerns in smart city deployments.
To achieve adaptability, it is crucial to understand and develop the supporting technologies and paradigms that underpin both the smart city and metaverse architectures. The architecture of the metaverse is multi-layered, with each layer supporting different functionalities crucial for the operation and expansion of the metaverse. In [30,40], the authors present the seven-layer architecture based on the concept of the metaverse in the context of the value chain [54], which includes the infrastructure layer, human-interface layer, decentralisation layer, spatial computing layer, economy layer, social layer, and experience layer. These layers map closely to a broader architecture comprising the human, physical infrastructure and digital spaces highlighted in [37]. Following the concept in [37], we adopt the architecture of the metaverse for smart cities, which consists of human interaction, the physical infrastructure, and the digital world. This architecture also integrates the four primary layers of the generic smart city architecture, creating a seamless interaction between these spaces. Figure 2 shows the generic architecture of the metaverse for smart cities, showing the connection between its various components.
The human interaction layer, represented by digital avatars and user interfaces, bridges the physical and digital realms through advanced human–computer interaction (HCI) technologies and extended reality (XR) devices like AR glasses and VR headsets [55]. These technologies enable citizens to engage in immersive social interactions and workspaces, and access services within the metaverse, enhancing their interaction with smart city services [55].
The physical infrastructure layer, corresponding to the sensing layer and part of the transmission layer in the smart city architecture, includes storage, computation frameworks, and sensor networks. These components ensure that massive data flows generated within the virtual environments are efficiently processed and transmitted in real time [56]. Cloud–edge-end computing infrastructures support this process, while high-speed networks like 5G/6G provide essential connectivity. This infrastructure is the backbone for integrating IoT devices that collect heterogeneous data from urban environments.
The digital world layer represents part of the application layer in smart city architecture and consists of interconnected virtual worlds or sub-metaverses [57]. This layer integrates data collected by IoT sensors to create immersive and functional virtual representations of urban scenarios through AI and digital twins. DTs mirror real-world environments, enabling applications like smart parking systems that provide real-time monitoring and navigation systems that enhance mobility with personalised, real-time guidance through augmented reality overlays and voice instructions.
The metaverse engine is responsible for creating and maintaining the digital world, using technologies such as AI, blockchain, and DTs to generate realistic virtual environments and facilitate trustworthy transactions [18]. This layer aligns with the data management layer of the smart city generic architecture. AI-based algorithms generate personalised content and avatars and provide intelligent services to enhance the metaverse ecosystem. This dynamic feedback loop between the physical and digital worlds ensures the real-time optimisation of city services.
In the metaverse, these services are further enhanced by providing a virtual, interconnected space where digital and physical realities converge, allowing for immersive, real-time interactions between city managers, citizens, and urban infrastructure. AI plays a role in enabling the metaverse to understand human language, recognise objects, learn from data, and make predictions through advanced techniques [22,53]. Additionally, AI-driven semantic communication systems optimise data transmission and resource usage, particularly for devices with limited capabilities, ensuring seamless and immersive user experiences in dynamic and resource-constrained environments [58]. In the next section, we delve deeper into these AI techniques and technologies, exploring their specific roles and applications within the metaverse for smart cities.
3. AI-Enabled Metaverse for Smart City
Artificial Intelligence (AI) is the simulation of human intelligence in machines programmed to think, learn, and make decisions in ways that mimic human cognitive functions. It encompasses various techniques and technologies, including ML, NLP, computer vision, and robotics, to enable machines to perform tasks that typically require human intelligence, such as problem-solving, reasoning, and language understanding [59]. This section delves into AI techniques and technologies, exploring their specific roles and applications within the metaverse for smart cities.
3.1. Advancements in AI Techniques from Machine Learning to Generative AI
AI techniques have facilitated seamless interaction between virtual environments and citizens in smart cities. Semantic learning, neuromorphic computing, and NLP are key AI technologies enabling the metaverse to understand and respond to human languages and actions in real time [22]. Semantic learning focuses on understanding and processing data by capturing the meaning and context rather than just patterns. Neuromorphic computing, inspired by the structure and functioning of the human brain, employs neural networks and hardware to mimic neural activity. It is designed for energy-efficient processing, enabling tasks like pattern recognition, sensory processing, and real-time decision making, particularly in edge and IoT devices [60]. NLP is instrumental in generating meaningful human–computer interactions by processing large volumes of natural language data, including speech and text, thereby enhancing virtual experiences.
Machine learning, a subset of AI, allows these technologies to learn from data and improve task performance without explicit programming [39]. It encompasses deep learning (DL) [61], reinforcement learning (RL), and learning paradigms such as supervised, unsupervised, and self-supervised learning. These techniques excel in identifying patterns and making predictions across domains such as image recognition, speech analysis, and predictive analytics, which are essential in metaverse smart city applications. Deep learning employs artificial neural networks (ANNs) with multiple interconnected layers to process and hierarchically understand complex patterns in the data [62]. Key techniques, especially deep neural networks (DNNs), include convolutional neural networks (CNNs) [63] and Recurrent Neural Networks (RNNs) [64,65]. Convolutional neural networks specialise in spatial data processing, particularly image and video data, while RNNs are designed to handle sequential data, making them ideal for time-dependent scenarios.
Reinforcement learning (RL) involves agents learning optimal strategies through trial-and-error interactions within their environment to maximise cumulative rewards. Reinforcement learning techniques, including value-based, policy-based, and actor–critic methods, are particularly effective in dynamic scenarios where adaptive decision making is essential [66]. Particularly, reinforcement learning and self-supervised learning offer more scalable alternatives, aiming to make metaverse AI task-independent. For example, Meta AI has advanced self-supervised learning by creating the data2vec model, which combines multiple modalities like speech, vision, and NLP, enabling more adaptable and unified learning across different tasks [67]. Despite these advancements, self-supervised learning remains a work in progress, and further research is needed to realise its potential across different smart city metaverse applications. Similarly, continual learning (CL) has emerged as a novel machine learning paradigm with potential in smart cities [68]. This paradigm constantly updates models to adapt to changing environments, where the learning tasks, data, and distributions can vary over time. In Figure 3, we summarise the role of AI, ML, and DL in the metaverse for smart city applications.
Deep learning has made remarkable advancements. In particular, DNNs such as CNNs and RNNs have paved the way for more advanced models. For instance, advanced neural network architectures like transformers [69] have revolutionised AI by effectively learning long-range dependencies, outperforming traditional RNNs in tasks such as machine translation and text summarisation [70]. The advancement of these models has been partly enabled by transfer learning, which leverages pretrained knowledge from massive datasets, reducing the need for large labelled datasets [71]. This has led to more advanced and capable models, such as GAI models, including LLMs.
3.2. Generative AI for the Metaverse in Smart City
Generative AI refers to a class of AI models that are designed to learn the underlying patterns and distributions of the training data to generate new samples similar to the existing samples [72]. To achieve this, GAI models use a latent space to represent the hidden features of data, which is then mapped into actual data space through a generator function. The generator’s parameters are fine-tuned during training to ensure the generated samples closely resemble the actual data. The training process optimises the likelihood that the model is virtually indistinguishable from the actual samples by focusing on maximising the probability of generating the observed data samples [73]. LLMs, a subset of GAI, play a foundational role in the broader capabilities of GAI.
Examples of LLMs, such as ChatGPT and Bard, have significantly advanced NLP by enabling human-like text generation, multilingual support, and contextual understanding [74,75,76]. In the metaverse for smart cities, LLMs can revolutionise interactions by providing real-time multilingual communication, facilitating citizen engagement, and generating content for education and training [77,78]. In addition, LLMs can create interactive learning materials and enhance intelligent transportation systems, as seen in studies [79,80]. Furthermore, examples of GAI applications in the metaverse include text generation, image generation, audio generation, and video generation, all of which facilitate the creation of virtual environments in smart cities [81]. While LLMs focus on natural language tasks, GAI encompasses a broader range of generative capabilities across multiple modalities and can be categorised into unimodal and multimodal models, as shown in Figure 4, which shows the general architecture.
In the metaverse smart city applications, multimodal models can offer a transformative approach to managing and simulating urban systems by capturing the relationships between diverse data modalities. Unlike unimodal models, which focus on a single data type, multimodal models learn joint distributions across multiple modalities. This capability is critical for metaverse applications in smart cities, where complex and interconnected systems require the coherent integration of heterogeneous data sources. For instance, a multimodal model could integrate traffic sensor data, weather data, and public transit schedules with different modalities to optimise city-wide mobility in real time. While unimodal models excel at tasks within a single data type, they lack the versatility to address the multidimensional challenges of smart city metaverse environments that represent real-world data [82]. However, integrating information from multiple modalities is an active research area [73]. Multimodal learning primarily uses two architectures: early fusion, which combines data at the beginning of the model, and late fusion, which processes modalities separately before combining results [83].
Diverse and advanced architectures are tailored to handle specific generative tasks within and across these modalities. These include transformers, autoregressive models, Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and flow-based, and diffusion-based models.
-
(1). Transformers: Transformers are neural network architectures designed to process sequential data efficiently. They utilise a self-attention mechanism, allowing the model to weigh the importance of different parts of the input sequence relative to each other [69]. Transformers are the backbone of state-of-the-art models like Generative Pretrained Transformers (GPTs) [84,85], bidirectional encoder representations from transformers (BERTs) [86], and T5 [71], excelling in tasks such as text generation, image generation, language translation, and NLP. Additionally, transformers streamline the computation of relationships between distant positions, enabling better parallelisation, faster training, and improved performance over traditional neural networks [73].
-
(2). Autoregressive models: Autoregressive models generate outputs sequentially by predicting the next element based on previously generated elements. Autoregressive models predict each element in a sequence based on the entire history of previous elements. Examples include Seq2Seq models [87,88,89]. This makes them ideal for NLP tasks requiring coherent sequence understanding and generation, as well as for modelling time-series data dynamics [73].
-
(3). Variational Autoencoders (VAEs): VAEs are probabilistic models that encode input data into a latent space, ensuring the latent representations follow a Gaussian distribution [90,91,92]. They then decode these representations to reconstruct the input data. This process involves two key components: encoders and decoders. In the VAE framework, the encoder maps input data to a probability distribution in a latent space, while the decoder reconstructs the original data from this latent representation. Both components are implemented using neural networks and are trained jointly through variational inference, which minimises reconstruction and regularisation losses [93]. VAEs utilise a reparameterisation trick, enabling backpropagation by transforming samples from a standard normal distribution to align with the encoder’s output distribution.
-
(4). Generative Adversarial Networks (GANs): GANs comprise two neural networks: a generator and a discriminator, which compete in a game-theoretic setting [94,95,96]. The generator learns to produce realistic, high-quality data, such as images, text, or music, by capturing the underlying distribution of the training data. Meanwhile, the discriminator evaluates this generated data, distinguishing between real and fake samples [97]. Through an iterative process, the generator improves its capability to produce data that closely mimic real samples, while the discriminator simultaneously enhances its ability to identify generated data.
-
(5). Flow-based: Flow-based models are deterministic and invertible transformations that map between raw data and latent space. Unlike generative models like GANs or VAEs, which rely on latent variables, normalising flow models directly learn the mapping between two distributions by manipulating the Jacobian determinant [98]. They apply a sequence of invertible transformations to a simple probability distribution to model complex distributions using affine coupling layers in the encoder. The decoder function serves as the exact inverse of the encoder, ensuring computational efficiency and maintaining the model’s tractability [98].
-
(6). Diffusion models: Diffusion models generate data by progressively denoising a noisy sample until it resembles the target distribution [99,100]. These models simulate a diffusion process by iteratively adding noise to an initial image and then learning to reverse the process to produce realistic samples. Approaches like Noise-Contrastive Estimation (NCE) [101] and score-based generative modelling [102] are commonly used within this framework. Diffusion models generate complex and varied outputs by transforming data from a simple distribution to a target distribution over a series of steps. This methodology has been integral to developing advanced generative techniques like Denoising Score Matching (DSM) and diffusion probabilistic models [73]. Diffusion models produce high-quality outputs, but their iterative denoising process is computationally expensive, memory intensive, and a slower sampling process than the other generative models.
Similarly, we classify how these models are applied to perform various tasks, including text generation, image generation, audio generation, video generation, and 3D object rendering.
-
(1). Text generation: Text generation focuses on generating textual content within the metaverse. It includes applications such as chatbot conversations, virtual world descriptions, interactive narratives, and text-based interactions between users. Models used for text generation tasks can be classified based on the type of data they deal with, which can be unimodal or multimodal. Transformer models such as GPT, BERT, and T5 excel at this task in unimodal models where textual output is generated from text input data. This also makes them useful in machine translation, question-answering (QA), and summarisation tasks. While autoregressive models ensure high-quality outputs, their sequential nature can lead to slower inference times, and they tend to generate generic text. VAEs enable controlled outputs by manipulating the latent space in text generation, but they often struggle with capturing long-range dependencies in sequential data. In GAN-based models, RNNs are used in their generator, discriminator/encoder, and decoder to generate text. While GANs generate diverse content, they face challenges such as training instability. Even though diffusion models were initially developed for image generation, they have been adapted for text-generation tasks [103,104]. Descriptive textual content can also be generated from other modalities such as image, audio, or video. GAI models like VisualGPT [105] and ViT [106] use transformer architectures to generate textual descriptions from image features, while models such as VideoCLIP [107] capture temporal information from sequential frames of the video. Models like DeepSpeech [108] and wav2vec 2.0 [109] focus on transcribing speech into text using RNNs and a transformer-based architecture that learns representations from masked speech input in latent space [72].
-
(2). Image generation: This entails producing image content for the metaverse to create an immersive visual experience. It includes tasks like creating virtual environments, objects, avatars, textures, and other graphical aspects [81] like representing virtual cities, traffic symbols, buildings, green spaces, and medical imaging. Most of these visual models are based on CNNs, which excel at capturing spatial hierarchies and local patterns. Unimodal image models can be categorised into autoregressive, VAE, GAN, and flow-based and diffusion models [72]. Autoregressive models like PixelRNN [110] and PixelCNN [111] generate images pixel-by-pixel, capturing local dependencies and patterns and predicting each pixel’s value using the previously produced pixels at a slower rate. VAE-based models such as IntroVAE [112] and VQ-VAE-2 [113] encode images into a latent space with Gaussian distributions, enabling parallel image generation. However, the outputs may not be of good quality. CycleGAN [114] and StyleGAN [115] are GAN-based models capable of producing realistic and clear images but face training stability and output diversity challenges. Hybrid models like CVAE-GAN [116] combine VAEs and GANs for enhanced generation quality. Flow-based methods like Glow [117] and real NVP [118] excel in image generation, offering improved quantitative performance in log-likelihood metrics. However, flow-based models can be less expressive and struggle with capturing long-range dependencies. Denoising diffusion probabilistic models (DDPMs) [100] and denoising diffusion implicit models (DDIMs) [99] are highly effective at image generation but computationally intensive and require significant memory. Multimodal image generation models, which convert input text into images, accomplish this by analysing the semantic content of the text and generating corresponding visual representations. Many of these models utilise a dual-stage architecture that combines a text-understanding model with an image-generating model. Examples of such models include DALL-E [119] and DALL-E 2 [120], which use transformer-based architectures combined with VAEs and diffusion techniques to generate high-resolution images based on text descriptions. Unlike DALL-E, which focuses on image generation, Contrastive Language–Image Pretraining (CLIP) [121] excels at understanding relationships between text and images through paired data. By employing contrastive learning on large-scale image–text pairs, CLIP encodes both modalities into a shared embedding space, enabling tasks such as image classification, object detection, and zero-shot learning. Built on a dual-encoder architecture, CLIP uses separate encoders for images and text, effectively capturing crossmodal relationships and supporting diverse applications beyond traditional image analysis.
-
(3). Video generation: Video generation in the metaverse involves creating dynamic visual content to enhance storytelling, engagement, and urban simulations. Applications include generating animated scenes of traffic flow patterns, virtual tours of proposed urban projects, interactive videos illustrating waste management practices, and dynamic recreations of disaster scenarios for emergency response training. Video generative models are developed to address various video processing tasks, including video super-resolution, video inpainting, and video denoising [72]. For instance, the study [122] introduced a video diffusion model, extending standard image diffusion architectures to jointly train on image and video data, producing realistic video clips. Another notable example is a GAN-based model, MoCoGAN [123] which separates motion and content components to produce coherent and realistic video sequences. Multimodal models add another layer of sophistication by translating text inputs into multi-frame videos. For example, Phenaki [124] employs a bidirectional transformer to predict video tokens from text embeddings, ensuring temporal coherence, though it requires significant computational resources. Further, innovations like OpenAI’s Sora [125] demonstrate the ability to create realistic, dynamic video content directly from textual descriptions. Similarly to DALL-E 2, Sora operates semantically, interpreting textual prompts and synthesising video sequences that are coherent and visually aligned with the instructions. Alternatively, models like CogVideo [126] and Tune-A-Video [127] use text-to-image techniques to generate sequential frames, which are then connected or interpolated into a complete video sequence.
-
(4). Three-dimensional object rendering: This includes generating 3D object representations of urban infrastructure such as buildings, roads, and vehicles, as well as interactive elements like characters, public installations, and other elements that populate the virtual environment. Diffusion models, GANs, and neural radiance fields (NeRFs) have demonstrated exceptional potential in addressing the challenges of creating 3D objects and assets [30]. For instance, StyleGAN [128] has effectively generated high-fidelity virtual assets like human faces, enabling rapid content creation with minimal additional data [129]. Similarly, DALL-E 2 [119] and Google’s Imagen [130] leverage natural language prompts to generate personalised 2D content, laying the groundwork for 3D object creation through textual descriptions. Additionally, Magic3D [131] also extends these capabilities to 3D object creation by improving the diffusion DreamFusion [132] model. However, 2D image generation is insufficient for immersive metaverse environments where users interact with 3D objects. NVIDIA’s Instant NeRF [133], built on deep neural networks, represents a significant breakthrough, converting 2D images captured from multiple angles into detailed 3D scenes. Recent advancements have integrated components from GANs and diffusion models to enhance NeRFs [134,135,136]. For instance, the GD2-NeRF [136] framework incorporates GANs and diffusion models to improve detail generation and maintain 3D consistency without per-scene fine-tuning. Similarly, the DiSR-NeRF [135] model utilises diffusion guidance to achieve view-consistent super-resolution in NeRFs, resulting in more realistic and detailed scene renderings.
-
(5). Audio generation: Audio generation in the metaverse encompasses creating virtual soundscapes, such as traffic noise and public announcements, spatial audio cues for the visually impaired, voice communication, and audio alerts for disasters or emergencies in real time. It enhances immersion by simulating real-world sounds, supporting communication, and enabling personalised and interactive audio experiences, such as tours, and storytelling in the digital realm. Autoregressive models, such as WaveNet [137] and SampleRNN [138], generate new sounds from raw audio waveforms, excelling in tasks like music generation, speech synthesis, and audio editing. GAN-based models like GANSynth [139] and WaveGAN [140] are also widely used for creating high-quality audio. Some models, like ArchiSound [141], SpecGAN [142], and VAE-VC [143], first transform audio data into visual representations, such as spectrograms, allowing them to leverage image-based generative techniques like diffusion, GAN, and VAE models for advanced audio manipulation. Models such as AudioLM [144] and Tacotron [145] map textual input to natural-sounding speech, enabling realistic voice synthesis. AudioLM employs multi-stage transformer architectures for syntactically and semantically coherent speech extensions while Tacotron, a Seq2Seq model, generates mel-scale spectrograms converted to audio using vocoders like WaveNet. Additionally, voice bots, including popular products like Google Assistant [146], Amazon Alexa [147], Apple Siri [148], and Baidu’s DuerOS [149], enable real-time voice-based interactions in the metaverse, such as in-car navigation systems. They process spoken inputs using automatic speech recognition (ASR), natural language understanding (NLU), and natural language generation (NLG) techniques to respond with human-like audio outputs [150].
3.3. Transformative Power of AI in the Metaverse for Smart City
As the previous sections show, AI is a transformative force in the metaverse, enabling seamless integration between virtual and physical worlds while enhancing user experiences through advanced techniques. For instance, RL enables avatars to interactively guide participants in virtual environments, enhancing functionality and immersion [151]. Similarly, AI-driven systems in the metaverse facilitate realistic human–computer interactions, such as voice-activated navigation and avatar behaviour modelling, using hybrid human–AI interfaces to solve socially oriented problems [152]. Deep learning (DL) techniques, including CNNs and RNNs, are employed for high-speed image processing, supporting real-time rendering and improved image fidelity. Furthermore, AI addresses challenges like multilingual accessibility by leveraging ML-based tools for language translation, enabling inclusive and global participation in smart city applications. Through innovations like AI-powered non-player characters (NPCs), underpopulated metaverse environments can transform into dynamic, engaging spaces, fostering social interactions and cultural activities [153].
Generative AI further transforms the metaverse in autonomous driving, urban systems, and interactivity. For example, generative models are used to simulate conditioned traffic and driving data to enhance safety and efficiency [154]. CityGPT, a large language model trained on city-specific data, supports embodied agents for managing urban complexities like congestion and pollution, bridging technological advancements and practical implementation [155]. Additionally, context-aware pipelines like PICAN assist users in navigating the metaverse by providing personalised spatiotemporal recommendations, enhancing engagement and immersion [156].
3.4. AI Performance Evaluation in the Metaverse for Smart City
Evaluating the performance of AI models in the metaverse is crucial to ensuring their reliability and alignment with human expectations across diverse smart city applications. Quantitative and qualitative metrics are essential tools in assessing these models’ effectiveness. Quantitative metrics like Inception Score (IS), Fréchet Inception Distance (FID), and Structural Similarity Index Measure (SSIM) are commonly used for image processing tasks to evaluate the quality, diversity, and structural alignment of generated images [98]. GANs and diffusion models utilise these metrics to improve image realism and fidelity. For text-based tasks such as language translation and content generation, metrics like BLEU, F1 score, and METEOR measure accuracy, fluency, and alignment with ground truth [98].
In immersive environments like the metaverse, qualitative metrics also play a vital role, capturing subjective factors such as visual appeal, coherence, and user satisfaction. For instance, human perceptual metrics assess the alignment of AI outputs with human expectations, while feedback from users and experts ensures relevance and usability [98]. For specific applications like 3D generation, metrics such as Chamfer Distance and Mean per Joint Positioning Error (MPJPE) evaluate the structural similarity and pose accuracy of the generated objects, essential for realistic virtual interactions [132]. In speech and audio tasks, intelligibility and perceptual quality are measured using Short-Time Objective Intelligibility (STOI) and Perceptual Evaluation of Speech Quality (PESQ) to enhance user engagement [157]. By leveraging these quantitative and qualitative approaches, the evaluation frameworks can ensure that AI systems in the metaverse meet the technical and human-centred standards, fostering more adaptive and immersive smart city applications.
Further, human-centred metrics such as the Technology Acceptance Model (TAM) can evaluate user acceptance and engagement [158]. Urban studies and surveys targeting specialists, decision-makers, and citizens can validate these metrics and gather insights into practical challenges and opportunities.
3.5. AI-Enabled Technologies for the Metaverse in Smart Cities
Smart city applications are multifaceted, with multiple dimensions that combine technology, innovation, and knowledge. This section overviews advanced AI-enabled technologies in the metaverse for smart cities. It discusses key technologies such as big data, NLP, computer vision, blockchain, digital twins, IoT, edge AI, and 5G/6G in realising a ubiquitous metaverse in smart cities. A summary of these technologies enabled by AI is presented in Figure 5.
3.5.1. Big Data
Big data, characterised by its volume, velocity, variety, and veracity, is generated from various sources in smart cities, such as IoT devices, social media, sensors, and spatial data far exceeding traditional city data volumes. This often unstructured data covers key aspects of urban life, including transportation, energy, environment, and public safety, forming the foundational basis for improving city services [36]. Integrating AI with big data enables the real-time analysis, prediction, and management of urban systems, allowing for problem anticipation, optimised resource allocation, and customised services [159]. In the metaverse, this integration becomes more complex, relying heavily on spatial and spatiotemporal data to blend physical and digital spaces, which is critical for realistic and immersive virtual experiences. AI techniques like spatiotemporal autoregressive models, time-series clustering, and graph-based models such as Graph Neural Networks (GNNs) and Graph Attention Networks (GATs) are employed to model complex dependencies and patterns [22]. Applications include traffic flow prediction under uncertain conditions [160], dynamic traffic management using real-time data from sensors and connected vehicles [161], and enabling the simulation of environmental conditions like weather patterns and natural disasters in the metaverse [162]. For instance, emergency response teams can practice in virtual disaster scenarios that change and develop over time [163], aiding urban planners in visualising the impact of various strategies. Furthermore, spatial–temporal models have been employed to recreate historical or real-world locations in the metaverse, allowing users to explore different periods or witness changes in a location over time, offering immersive educational experiences [164].
3.5.2. Natural Language Processing
Natural language processing, a key aspect of AI, enables the metaverse to understand and interpret human languages, facilitating seamless interactions between users and virtual environments, thus enhancing user experiences with more natural and personalised engagement with virtual agents [22]. Recent advancements in AI, such as LLMs, have significantly propelled NLP capabilities, enabling sophisticated conversation design and language modelling essential for immersive and interactive experiences in smart cities [165]. Advanced neural network architectures, including RNNs, LSTMs with attention mechanisms, and CNNs, along with hybrid models like LSTM-CNN, have been employed to improve the reliability and performance of virtual assistants [181 and generate short and long text for image captioning [40]. Furthermore, AI-based agents can leverage LLMs and open geospatial data to assist users in navigating complex digital environments, answering questions, and performing tasks [166] by utilising knowledge graphs and recommender systems for personalised support in urban systems. NLP-driven chatbots and dialogue systems also facilitate social interactions and collaborations within the metaverse, simulating realistic conversations [167] and enabling expressive communication through avatars. Further, deep learning-driven semantic communication systems utilise CNNs and RNNs to extract text-related semantic features from input speech, reducing data transmission and optimising communication resources [168], while NLP-based semantic communication systems for text transmission focus on restoring the meaning of sentences, unlike traditional communication systems.
3.5.3. Computer Vision
Computer vision consists of various stages—including image acquisition, preprocessing, feature extraction, object recognition, tracking, and interpretation—that allow machines to understand visual inputs, crucial for developing systems in the metaverse [169]. CV enables extended reality (XR) devices to interpret and comprehend user actions by processing visually meaningful data [40], allowing users, represented as avatars, to navigate 3D spaces and engage with virtual objects in the metaverse. CNNs have revolutionised CV by mimicking the human visual system and achieving remarkable accuracy in object and pattern recognition [170]. This has led to applications in the metaverse for creating realistic avatars, 3D spaces, and digital twins that mirror real-world objects and environments, such as ZooBuilder for animating animals using advanced pose estimation techniques [171]. Generative AI, including GANs and diffusion models, are crucial for creating diverse and high-quality 3D objects for the metaverse in smart cities. For instance, StyleGAN has been used for image-to-image translation, generating high-fidelity human faces without additional facial data [137], while tools like DALL-E 2 generate 2D images from text descriptions, as discussed in Section 3.2. An example using DALL-E 2 is shown in Figure 6. CV is also evident in designing and developing 3D spaces and virtual worlds in smart cities, allowing users to interact with virtual objects and each other in a three-dimensional environment. Researchers are optimising these spaces with virtual reality architectures to enhance human cognition [172] and collision detection algorithms for building placement [172]. Other smart city applications enabled by CV technologies include monitoring urban environments [173], enhancing public safety [174], and facilitating intelligent transportation systems [154,175]. Further, MR holds greater potential for the metaverse in smart cities due to its blend of physical and virtual experiences, facilitated by holographic devices with see-through displays and immersive devices that allow interaction with virtual objects.
3.5.4. Digital Twins
Digital twin technology is a virtual representation of physical objects, systems, or processes that facilitates real-time analysis, learning, and reasoning [173], synchronising real-world objects with their virtual counterparts for real-time monitoring, predictive analysis, and remote operation through AI and machine learning algorithms. In the metaverse, DTs manage interactions between AR/VR devices, offering advanced perception and cognition capabilities by using multimodal models, knowledge graphs, and IoT-based systems to create a comprehensive understanding of device-environment interactions. They provide a unified semantic mapping of these interactions, linking entity attributes and relationship knowledge [176], with knowledge graphs offering structured knowledge and connecting key concepts in the metaverse. In addition, DT models enhance connectivity and interoperability across devices using edge processing [31]. In smart cities, DTs enhance environmental sustainability by optimising space utilisation, increasing human productivity, improving maintenance efficiency, and reducing carbon emissions in building operations [177]. They also optimise traffic flow, improve public transportation systems, facilitate urban planning [178], and minimise downtime and costs by anticipating potential issues before they arise [179]. DTs monitor energy consumption in buildings by analysing interactions between energy sources, distribution networks, and consumers, facilitating efficient energy-saving strategies and the integration of renewable energy sources [180]. Urban digital twins allow for multisystem simulations and visualisations, enabling city planners to model buildings and environmental factors digitally [30]. In industrial manufacturing, DT is used to model production processes, predict equipment failures, and plan maintenance proactively, thereby increasing productivity and reducing waste [181]. While DTs primarily enhance realism and facilitate advanced services and social interactions within the metaverse, they are also crucial for integrating IoT networks that provide real-time data, a task that AI and machine learning continue to improve [31].
3.5.5. Blockchain
Blockchain technology enables the metaverse by providing a decentralised, transparent, and secure digital ledger for various operations [36] and has evolved into a versatile tool in domains like healthcare, real estate, and smart cities [36]. Fusing AI with blockchain is critical for overcoming challenges in big data analytics, AI-enabled content creation, and intelligence deployment, which are essential for evolving digital economies within virtual environments [182]. AI enhances blockchain’s capabilities in managing high transaction volumes typical of virtual environments, where digital assets like NFTs enable avatars to generate tradable content [183], often surpassing the scale of physical world transactions. In secure computing environments, blockchain and AI facilitate distributed data storage and enhance security and privacy within smart city infrastructures by addressing the vulnerabilities of traditional centralised systems [36]. For instance, Woods et al. [184] explored the fusion of AI techniques with blockchain infrastructure to address security risks in the metaverse, and Liu et al. [185] proposed a blockchain-based spatial crowdsourcing (SC) system (BlockSC) for the metaverse, addressing the challenges of centralised authority reliability and location privacy in conventional SC systems illustrated in Figure 7. In edge computing-based smart grids, blockchain facilitates secure transactions between energy suppliers and consumers at the network edge, though challenges such as optimising consensus among nodes persist [186]. Blockchain’s application extends to smart cities, particularly in Building Information Modelling (BIM) projects [187], enhancing the security of digital mapping processes in IoT, and improving the data reliability of digital twins. However, integrating these technologies poses significant challenges, such as the need for advanced AI techniques to manage complex interactions and high traffic volumes from numerous avatars and digital entities [184], and the fundamentally different economic structures within the metaverse necessitating innovative approaches to economic management [188].
3.5.6. Internet of Things
Internet of Things and AI technologies are crucial in enabling immersive digital experiences within the metaverse by mapping real-time IoT data from the physical world into virtual environments in the form of DTs [31]. This enhances AR/VR interfaces through IoT-enabled devices, allowing the real-time control of physical or virtual objects [31] and improving context and situational awareness by facilitating data exchange between the digital and physical worlds [189]. In addition, integrating semantic techniques with IoT, such as semantic communication, enables seamless interaction between real-world data and virtual environments, enhancing the functionality of AR/VR applications by addressing the variability in object meanings depending on the context [22]. AI-driven semantic technologies play a crucial role in processing complex data within IoT ecosystems, supporting interoperability and contextual information provisioning through a semantic layer that allows IoT devices to communicate despite differences in syntaxes or protocols [190]. Advanced wireless technologies like 5G and 6G further support semantic IoT by enabling resource-intensive semantic processing tasks to be transferred to federated edge servers, optimising semantic-aware networking [191]. In smart cities, IoT systems can standardise and fuse diverse urban data through semantic information, improving resource usage and decision-making processes [22]. Semantic technologies enhance data value by modelling and annotating various data types [192], simplifying complex urban environments and improving the efficiency of city services. Practical applications include managing crowd movement and mobility in cities like Gold Coast, Australia, and Santander, Spain [193].
3.5.7. Edge AI and 5G/6G Communication
Edge AI enhance the metaverse by enabling low-latency, high-performance computing at the network’s edge, essential for immersive VR/AR activities [192]. Various AI techniques, including soft computing, have been applied to address challenges at the network’s edge, such as active user detection [194], resource allocation [195], scheduling [196], and power management [197]. Additionally, DTs facilitate real-time synchronisation between the physical and virtual worlds using distributed edge servers and devices for sensing and actuation [37], as illustrated in Figure 8, enabling users to manage and control physical devices through DTs, including in space–air–ground integrated networks (SAGINs) [198].
High-speed data communication supports connected AR/VR experiences and real-time services requiring ultra-low latency due to haptic signals and human perception. Fifth-generation technologies like mmWave, NOMA, and massive MIMO enhance network capacity, while 6G is expected to connect individuals, IoT devices, autonomous vehicles [199,200], wearable sensors, and mobile robots [201], supporting smart city infrastructure and security applications, including the Internet of Vehicles (IoVs) [202]. AI is also utilised in spectrum management [203] to maintain ultra-reliable low-latency communication (uRLLC) standards, which ensures dependable data transmission with minimal delay for real-time applications. However, challenges like ensuring consistent global coverage and addressing physical latency limitations persist, highlighting the ongoing need for AI-driven innovations to fully realise the potential of the metaverse in future 6G networks [30].
Table 3 summarises the key AI-enabled technologies used in metaverse-based smart cities, provides examples of corresponding AI techniques applied, and outlines how these techniques and technologies address some challenges in such environments and general applications.
4. Metaverse Applications in Smart City Services
In this section, we discuss the application and use cases of the metaverse for sustainable smart cities, summarised in Figure 9, and some metaverse implementations in real-world smart cities.
4.1. Applications and Use Cases
Various literature sources offer varied interpretations of smart cities, which can be defined by the following domains: smart environment, smart mobility, smart energy, smart health, smart governance, and smart economy. We consider these domains in our discussion of the applications and use cases for sustainable smart cities due to their relevance to sustainability. These applications are summarised in Table 4.
4.1.1. Smart Environment
A smart environment encompasses a range of services to manage pollution-related issues, including air quality, waste management, water pollution, radiation pollution, agricultural systems, and ocean environment monitoring [6,204,205]. It aims to enhance the quality of life while minimising environmental footprints through technology, contributing to sustainable urban living. The metaverse plays a transformative role in creating smart environments for smart cities by using virtual simulations to assess the consequences of urban projects, such as new constructions or changes in land use, thereby promoting sustainable growth and reducing pollution. For instance, the simulation of various environmental conditions, such as changing weather patterns, day–night cycles, or natural disasters [15,206,207], is made possible. Several studies have explored the integration of advanced AI techniques in environmental applications within smart cities, particularly in the areas of air quality control [208,209], pollution classification [210], waste management [211,212], building management [213,214], and remote sensing [215] for monitoring and managing urban environments.
Table 4A summary of AI-enabled applications and use cases in the metaverse across various domains in smart cities.
| Domain | Sub Domain | References |
|---|---|---|
| Smart Environment | Air quality control | [208,209,216] |
| Environmental monitoring | [210,215] | |
| Disaster planning | [15,206,207] | |
| Water management | [217] | |
| Agriculture | [218,219,220] | |
| Waste management | [205,211,212,221] | |
| Infrastructure management | [213,214] | |
| Smart Mobility | Traffic flow prediction | [160,222,223,224,225] |
| Traffic monitoring | [226,227] | |
| Traffic condition analysis | [228,229,230,231] | |
| Autonomous vehicles and predictive maintenance | [154,232,233,234,235,236,237] | |
| Smart Energy | Energy management and forecasting | [238,239] |
| Energy optimisation | [214,240,241] | |
| Power grid management and monitoring | [242,243,244,245] | |
| Smart Health | Pandemic forecasting | [246,247] |
| Public health | [248] | |
| Medical care and management | [249,250,251,252,253] | |
| Medical training | [254,255] | |
| Smart Governance | Electronic voting | [256] |
| Decision making and service delivery | [257,258,259,260] | |
| Smart Economy | Smart payment systems and e-business | [261] |
| Smart manufacturing | [262,263,264] | |
| Enhance customer experience | [265] |
Despite the central role of cost-effective IoT-based sensors, challenges in large-scale installation and maintenance have led to research on hybrid IoT frameworks and intelligent system designs for the feasible and cost-effective monitoring of environmental factors [6]. Spatiotemporal dynamics enable the real-time monitoring of adaptive urban systems and predictive analysis of environmental changes. For instance, it can facilitate intelligent analysis and prediction, improving the management of air quality, waste, water pollution, and climate change issues [266]. Further, AI-driven solutions address environmental inequalities in urban areas by using real-time air quality data to inform policies for vulnerable communities [29], contribute to water management through the efficient monitoring and prediction of water demand [217], and facilitate precision farming, disease detection, and resource management in agriculture [218,219,220], leading to more sustainable practices. In waste management, AI enhances the effectiveness of handling and sorting recyclables; for example, the RECLAIM (
4.1.2. Smart Mobility
Smart mobility relies on technology to develop transportation and logistics networks, including gas and electric vehicles, traditional vehicles, taxi tracking and sharing services (e.g., Didi, Uber, and Lyft), bike/scooter/bicycle tracking, rail systems, aiming to reduce congestion, improve safety, and benefit the economy. In the metaverse, AI enables smart mobility by predicting traffic flow [222,223], monitoring traffic [226,227], analysing traffic conditions [228,229,230], managing congestion [39], and developing autonomous vehicles [234,235,236]. Moreover, combined with DTs, it simulates real-world traffic conditions, improving public transportation routes and assisting urban planners, traffic engineers, and policy-makers in making better-informed decisions about urban planning [22]. The metaverse provides virtual testing environments where autonomous vehicles can be trained on extensive data, closely replicating real-world conditions [267]. AI models equip vehicles to detect and classify environmental elements for safe navigation, as demonstrated by Tizedes et al. [231], who developed a system combining AI and LiDAR for efficient road hazard detection. Predictive analytics, relying on AI models and algorithms, integrates historical traffic patterns, real-time vehicle flows, weather conditions, and social media trends to predict traffic behaviours and recommend efficient routes, optimising traffic flow and aiding emergency response [237]. Innovative strategies like those by Nuvvula et al. [232] optimise electric vehicle fleet operations by integrating predictive analytics with renewable energy sources, improving maintenance, charging schedules, and route planning, leading to reduced emissions and cost savings. Bellini et al. [233] presented a model that manages diverse mobility data to predict and simulate the impact of different scenarios on urban transportation, visualised using a 3D City Digital Twin on the Snap4City (
4.1.3. Smart Energy
Smart energy leverages technologies like AI, IoT, XR, smart grids, and renewable energy sources to create energy systems that monitor, predict, and adapt to changes in energy demand and supply, enhancing decision making and enabling proactive responses in urban infrastructure. In the metaverse, these systems support the seamless integration of renewable energy sources, intelligent storage solutions, and decentralised microgrids, crucial for energy stability, reliability, and sustainability. AI enhances the efficiency and reliability of renewable energy by enabling the accurate forecasting of supply, energy, and grid management [239,244,245], and, when integrated with digital twins (DTs), can model interactions between energy sources, distribution networks, and consumers to optimise energy consumption in buildings and across cities by implementing energy-saving measures and integrating renewable sources [214,241]. AI combined with XR and IoT improves the maintenance of geographically dispersed generation assets through cost-effective remote assistance and virtual site inspections, enhancing workforce productivity and operational efficiency via advanced data analytics and real-time monitoring [37]. For instance, an IoT-based framework integrates wind and solar data using a hybrid model combining a multi-head attention-based deep autoencoder with an extreme gradient boosting algorithm for accurate power generation forecasting in smart microgrids [242]. Similarly, He et al. [243] introduced a data-driven situation awareness paradigm (DT-SA) utilising a digital twin model for the Energy Internet of Things (EIoT-DT), improving system performance through continuous virtual-real interaction and feedback, with case studies including smart city EIoT-DT construction in Lingang, China, and power grid monitoring. Additionally, integrating IoT and smart grid technologies in residential settings significantly enhances energy efficiency and management, balancing energy demand and supply to create sustainable urban environments [240].
4.1.4. Smart Health
Smart health leverages wearable devices, IoT, and wireless communication technologies to manage health information, connect individuals and resources, and efficiently respond to health environment demands [39]. In the metaverse for smart cities, AI, wearables, and the Internet of Nano-Things (IoNT) help track infectious diseases by leveraging real-time medical data from large populations [246], enabling the early detection of outbreaks and allowing healthcare providers to prevent disease spread swiftly [268]. For instance, Farooq et al. [246] introduced a DNN-based, data stream-driven real-time online incremental learning algorithm to analyse COVID-19 transmission dynamics and prevention strategies, offering insights for pandemic forecasting and policy recommendations. Haptic technologies enable physicians to remotely interact with patients, benefiting high-risk groups like healthcare workers during pandemics. Additionally, 6G technology enhances emergency response efforts by improving positioning accuracy, allowing rescue teams to locate individuals more effectively during natural disasters like wildfires and earthquakes [247,248]. Integrating AI and holographic technology in ambulance services connects hospitals and ambulances, enabling doctors to conduct rapid patient assessments with paramedic support. The metaverse facilitates a shift to immersive 3D virtual care, offering advanced applications such as surgery assistance, health management, and real-time patient monitoring in a more interactive environment [37]. Additionally, AI enhances Internet of Medical Things (IoMT) devices by improving diagnostics and enabling the continuous monitoring of health conditions like heart rate, blood glucose levels [39], and chronic illnesses [250,251,252]. Further, AI-enabled healthcare simulations within the metaverse provide safe environments for medical professionals to practice skills through realistic scenarios [267,268] and aid in training patients with disabilities [249]. Moreover, AI-powered virtual assistants like chatbots within the metaverse help users navigate healthcare services and answer health-related questions [253], enhancing user experience. The metaverse also enables AR/VR therapy sessions in therapeutic environments, helping patients manage stress, anxiety, or other mental health conditions [22].
4.1.5. Smart Governance
Smart governance leverages technologies to enhance decision making, policy development, and service delivery by facilitating information exchange, communication, and collaboration with stakeholders through various ICT applications like social media, blogs, and public forums to improve government services [6]. Integrating AI with social innovation (SI) significantly enhances smart decision making in governance, as seen in South Korea and Pakistan [269]. Decision Support Systems (DSSs) play a crucial role in promoting inclusiveness in smart city design, ensuring services are accessible to all the citizens, particularly the vulnerable, as demonstrated by Abu Qasem et al. [257] in the Ageing Smart (
4.1.6. Smart Economy
The concept of a smart economy involves the integration of advanced technologies such as AI, blockchain, and big data analytics in areas like e-business and e-commerce, including manufacturing and service delivery, to create innovative and sustainable economic models. In the metaverse for a smart city, these technologies optimise business processes, enhance financial transactions through smart payment systems and blockchain, and drive new industries focused on sustainability, such as green energy and circular and sharing economy [39]. For instance, Abraham et al. [261] developed a smart toll transaction application with an AI-enabled multi-agent system and blockchain for dynamic pricing and traffic optimisation for smart transportation systems in smart cities. The metaverse in a smart city offers unique marketing opportunities where brands can interact with users in immersive virtual environments, which integrates real-time advertising within gaming [30] and social media platforms. In addition, one assumption about the metaverse is the possibility of a creator economy, where trading assets across different spaces is facilitated [271]. AI further enhances these applications by providing personalised recommendations, virtual assistants, and chatbots that mimic human interactions, creating a more engaging and tailored user experience [272]. Additionally, in a smart economy, the metaverse supports entrepreneurship by establishing new business models, which reduce physical store costs, streamline supply chains, and enhance customer interaction through immersive experiences [273]. Kim et al. [265] studied how AR and VR technologies affect consumer perceptions and behaviours in virtual shopping environments by providing realistic experiences like simulating the touch of fabric and the smell of perfume. In smart manufacturing, Sun et al. [263,264] introduced a learning-based framework designed to help acquire complex 3C assembly skills using a multimodal digital twin environment, and Hou et al. [263] developed a hybrid residual multi-expert reinforcement learning method for interactive learning in the digital industrial metaverse.
4.2. Existing/Ongoing Sustainable Applications in Real-World Metaverse Smart Cities
From Singapore to the USA, cities worldwide are slowly adopting the metaverse to help them monitor and control groundwater, manage urban services, reduce congestion, prepare for emergencies, and optimise energy effectively. We give a few examples of current and ongoing metaverse use cases in real-world smart cities.
4.2.1. Virtual Singapore, Singapore
Singapore continues to lead in metaverse-related innovations, leveraging DT and GAI to enhance its digital infrastructure and urban management. Building on its position as the first nation to develop a comprehensive digital twin, the city-state has introduced over 100 GAI solutions under a government initiative, including chatbots designed to facilitate community centre operations [274]. As a digital twin of the city-state, Virtual Singapore integrates advanced 3D mapping and geospatial technologies, providing a dynamic, interactive model of the urban environment. Developed from the Singapore Advanced Map, which involved data collection through aerial photography, LiDAR, and mobile mapping, the platform enables stakeholders across government, business, and academia to test solutions, plan infrastructure, and conduct research in a risk-free virtual environment [259]. Further, the Digital Urban Climate Twin (DUCT) employs advanced computational modelling, data analytics, and simulations to analyse urban climate and environmental conditions [275]. This project facilitates collaboration among government agencies, researchers, and industry partners, enabling informed urban planning and sustainability decisions. Despite challenges such as high costs for data collection, infrastructure, and maintenance, as well as ensuring data interoperability and user accessibility, Singapore’s strategic use of open standards and collaborative projects minimises hurdles.
4.2.2. Robotaxi, China
China is at the forefront of integrating autonomous vehicles into its smart cities, with major cities like Beijing, Shanghai, and Guangzhou serving as testing grounds for robotaxis [276]. Companies such as Baidu, Pony.ai, Didi, and WeRide are competing to dominate this emerging market, leveraging their technological advancements to offer innovative transportation solutions. These efforts align with the nation’s goal of creating connected smart cities, where autonomous vehicles contribute to reduced traffic congestion, enhanced road safety, and transformative changes in urban mobility systems. Robotaxis, projected to hold a 5% market share of shared mobility kilometres in China by 2030, are anticipated to revolutionise urban transportation by reducing reliance on traditional ride-hailing and car-sharing services. Baidu’s Apollo Go platform has already provided over 7 million rides, showcasing advancements in autonomous driving safety and operational efficiency. Other players like Pony.ai are scaling up their fleets and expanding internationally. Despite these advancements, challenges persist, including concerns over safety, customer trust, and regulatory issues.
4.2.3. Metaverse Seoul, South Korea
Seoul has emerged as a pioneer in urban innovation with the launch of Metaverse Seoul [260], the world’s first municipal metaverse platform. This groundbreaking initiative enhances urban life through a virtual environment that blends technology with civic services, focusing on three core values: freedom, inclusion, and connection. Users can create personalised avatars and access a range of public services, from visiting the virtual mayor’s office, filing civil complaints, issuing official documents, paying local taxes, participating in public contests, and emphasising creative expression and accessibility. The platform also promotes inclusivity by enabling barrier-free interactions for all users, allowing access to facilities like the Seoul Fintech Lab and virtual tours of the city’s top landmarks, including N Seoul Tower and Gyeongbokgung Palace. This effort positions Seoul as a prototype for municipal metaverse administration and aligns with sustainability goals by reducing physical commutes and energy consumption. Metaverse Seoul sets a bold vision for the future of urban living, blending physical and virtual spaces to redefine connectivity and civic engagement.
4.2.4. SmartWorldPro—Energy, New Mexico, USA
Smart energy systems are another meaningful use of DT, essential for decarbonisation and the transition to renewable energy. New Mexico, the U.S.’s fifth largest state, worked with Cityzenith and its platform SmartWorldPro to create modelling for smart infrastructure to reduce carbon emissions by at least 50%. SmartWorldPro facilitates high-resolution 3D digital twin modelling for the efficient planning, design, and operation of new infrastructure [277]. The platform integrates diverse data and software, enabling the seamless optimisation of smart cities, energy networks, and urban planning. With applications extending to sectors like healthcare, transportation, and public utilities, New Mexico’s project also aims to establish six new smart cities and significantly improve residents’ quality of life through a Living-as-a-Service feature.
4.2.5. Flood Resilience, Lisbon, Portugal
Lisbon is leveraging metaverse technology to address the increasing flood risks posed by climate change and rapid urbanisation [278]. This innovative approach integrates real-time spatial data from drones, satellite imagery, and IoT sensors to create a highly detailed and dynamic 3D model of the city’s infrastructure. The DT flood resilience model enables authorities to simulate stormwater flows, assess vulnerabilities, and optimise drainage systems. By dynamically linking these models to the existing stormwater and sewer networks, Lisbon can visualise and simulate the impacts of extreme weather events, including storm surges and heavy rainfall, on critical urban assets.
5. Discussion
Generative AI emerges as a cornerstone in advancing the metaverse for smart cities, building upon foundational AI techniques such as ML and DL. GAI models like GANs and diffusion models have significantly expanded the scope of metaverse applications, including AR, VR, and XR [72]. For example, GANs are instrumental in creating high-fidelity content such as lifelike avatars, virtual buildings, and dynamic cityscapes, which are essential for visually appealing and interactive virtual environments [279]. Meanwhile, diffusion models excel at generating detailed images and videos, enabling dynamic virtual scenarios like traffic simulations [280] and disaster response training critical for real-time decision making in smart cities.
The digital twin, increasingly recognised as a practical tool for managing interactions between AR/VR devices in the metaverse, plays a pivotal role due to its perception and cognition capabilities. In [178], a smart city digital twin architecture is introduced to facilitate semantic knowledge representation and reasoning. By integrating GAI-driven semantic communication, digital twins can act as intelligent intermediaries, enabling the real-time simulation and analysis of interactions while bridging physical and virtual environments [72]. This enhances interoperability and coordination among AR/VR devices, supporting advanced decision making by interpreting contextual information and adapting to the dynamic conditions typical of smart cities.
Digital twins further enable the real-time monitoring, forecasting, and optimisation of urban resources, including traffic management, energy consumption, and waste generation [22]. Their scalability and adaptability are evidenced by successful implementations such as Virtual Singapore [259,275] and Metaverse Seoul [260], which showcase how smart city digital twins can drive sustainable and efficient urban management solutions.
However, creating a seamless integration between the physical and virtual worlds requires vast amounts of multimodal data, including text, images, and videos. Recent research has begun addressing this challenge through data augmentation techniques that leverage advanced frameworks such as diffusion models [280] and GANs [281] to generate high-quality digital content for smart city applications in intelligent transportation.
User experience quality is another critical consideration in metaverse applications for smart cities. As these environments grow in complexity and functionality, it becomes paramount to design GAI-driven systems capable of delivering personalised services tailored to the unique needs of individual users [72]. By doing so, researchers and developers can ensure that the metaverse continues to provide value to diverse stakeholders while supporting the broader goals of smart cities.
Nevertheless, given the evidence of these technologies’ transformative opportunities, they also bring significant challenges that must be addressed. This section acknowledges the limitations of current AI models and outlines future research directions.
5.1. Challenges in Deploying AI-Enabled Metaverse for Smart City
Difficulty in obtaining high-quality training data presents a significant hurdle. The metaverse’s unique and dynamic environment necessitates collecting high-quality, diverse, and representative data to train effective generative AI models. However, the complexity and variability inherent in virtual urban settings make data acquisition difficult. Limited or biassed data can compromise the performance and realism of AI-generated content, diminishing user experiences and undermining the metaverse’s overall functionality within smart cities [281].
Generating realistic and high-fidelity content is essential for creating immersive and engaging virtual environments. Achieving the required level of realism demands sophisticated generative AI technologies capable of producing detailed and authentic virtual elements [81]. Without such high-fidelity content, virtual spaces may appear artificial. This leads to user disengagement and reduces the metaverse’s appeal and utility in smart city applications like virtual tourism, urban planning simulations, and interactive public services.
The integration of GAI also brings forth ethical risks. On one hand, GAI facilitates personalised content creation, enhancing user experiences and fostering creative expression. On the other hand, it poses risks such as spreading misinformation and deepfakes [81]. The ability to generate misleading information can erode trust within the metaverse, making users susceptible to manipulation. Additionally, GAI models trained on biassed data can perpetuate and amplify the existing societal biases, resulting in the unfair treatment of certain user groups and contributing to a hostile or unwelcoming virtual environment [73]. Furthermore, advanced AI techniques, such as LLMs, often lack sufficient consistency for actionable decision making and are prone to hallucinations when trained on insufficient data [73]. This limitation poses risks in critical smart city applications where accurate and reliable information is essential for public safety and welfare.
Privacy, consent, and ownership concerns further complicate the deployment of GAI in the metaverse. GAI systems may inadvertently collect and utilise personal data without explicit consent, leading to breaches of user privacy and a loss of trust in the platform [73]. Moreover, determining the ownership and rights of AI-generated content is inherently complex, given the collaborative nature of creation between humans and machines. Ambiguities in ownership can result in legal disputes and hinder the commercialisation and protection of digital assets, necessitating robust privacy safeguards and clear legal frameworks [81].
Balancing innovation with responsible use is crucial to ensure the sustainable growth of the metaverse. While leveraging generative AI can drive significant advancements and innovations in smart city applications, it is imperative to implement guidelines and oversight mechanisms that promote responsible use [73]. Without this balance, the rapid deployment of generative technologies may lead to ethical and regulatory dilemmas that could threaten the long-term sustainability of the metaverse, impeding its role in enhancing urban living.
Technical limitations and scalability pose additional challenges. The current generative AI technologies may struggle with scalability and efficiency when applied to the expansive and intricate environments of the metaverse in smart cities. Technical bottlenecks, such as computational constraints and the inability to process large-scale data in real time [73], can restrict the development of seamless and expansive virtual worlds. These limitations hinder the metaverse’s potential to provide integrated and interactive urban experiences, thereby limiting user engagement and the effectiveness of smart city solutions.
The regulatory and legal frameworks necessary for governing generative AI in the metaverse are currently inadequate. The absence of comprehensive regulations specific to generative AI in virtual environments leads to inconsistent practices and legal ambiguities [81]. This regulatory gap complicates the enforcement of ethical standards and the protection of user rights, creating an environment where misuse and unethical practices can proliferate unchecked. Establishing tailored regulations and legal standards is essential to ensure that technological advancements align with societal values and legal norms.
Another challenge lies in performance measurement and evaluation. Establishing consistent evaluation metrics and benchmarks is necessary for assessing the effectiveness of AI models in metaverse applications across various smart city domains. Quantitative and qualitative metrics must be developed to evaluate technical performance and human-centred aspects like user satisfaction and engagement [98,158]. The complexity and diversity of metaverse applications in smart cities require comprehensive evaluation frameworks to ensure reliability and alignment with human expectations.
Lastly, the interoperability and standardisation of AI systems and content across different metaverse platforms present significant challenges. The metaverse comprises diverse virtual environments and services, each potentially utilising different architectures and standards [81]. Integrating these systems becomes difficult without common standards, leading to fragmented user experiences and reduced platform cohesion. Addressing interoperability issues is essential to facilitate the seamless movement of users and assets across various virtual spaces, ensuring the metaverse functions as a unified and interconnected digital ecosystem.
5.2. Future Research Directions
Building upon the identified challenges, several future directions are pivotal to advancing AI in the metaverse for smart cities. Addressing these areas will create more immersive, realistic, and ethically responsible virtual environments.
5.2.1. Advanced Data Collection and Augmentation
To overcome the difficulty in obtaining high-quality training data, future research should focus on developing advanced data collection techniques tailored to the metaverse’s unique environment. This includes leveraging sensors and interactive platforms to gather diverse data types. Additionally, advanced data augmentation methods as proposed in [280,281] can enhance the quality and diversity of the existing datasets. Synthetic datasets that accurately mimic the smart city characteristics should also be prioritised, utilising simulation tools and algorithms to generate realistic virtual scenarios for training.
5.2.2. Advancing Model Architectures and Training Algorithms
Research should focus on improving model architectures, training algorithms, and unsupervised learning techniques to enhance the performance and reliability of GAIs and LLMs [73,81]. This includes developing methods to mitigate hallucinations, biases, and lack of explainability, and ensuring that AI systems are reliable, transparent, and trustworthy. Further, research should consider adapting these models to various domains in smart cities.
5.2.3. Explainability and Responsible AI
Developing explainable AI and responsible AI techniques is crucial to bridging the gap between performance and interpretability. Understanding AI models’ decision-making processes ensures transparency, fairness, accountability, and trustworthiness in AI-driven decisions in the metaverse for smart cities [282,283]. Future research must address the challenges related to preventing bias in AI algorithms, ensuring fairness, and verifying reliability without solely relying on their decision-making processes.
5.2.4. Enhancing Security and Privacy Measures
Implementing robust encryption methods, secure authentication protocols, and privacy-preserving techniques such as data anonymisation and decentralised identity management is essential [22]. Federated learning can play a critical role by training ML models on decentralised data sources without sharing raw data, thus maintaining privacy. Ensuring long-term sustainability requires securing metaverse interfaces against cyber-attacks and safeguarding sensitive user information [31].
5.2.5. Ethical and Regulatory Frameworks
Developing robust regulatory frameworks is necessary to align virtual interactions with real-world regulations and social norms. Ensuring user data privacy while maintaining transparency in data collection, storage, and utilisation is essential [22]. Effective measures and policies must protect users from harmful behaviours such as cyberbullying, harassment, and abusive language [30]. Efforts should promote cultural diversity and representation within virtual environments, and AI systems must minimise biases to prevent discrimination or inequality.
5.2.6. Interoperability and Standardisation
Ensuring interoperability and standardisation across different metaverse platforms will facilitate the seamless integration of AI systems. To enable interoperability between diverse metaverses, research should focus on developing cross-domain semantic mapping and alignment techniques for heterogeneous ontologies, vocabularies, and data representations while addressing inconsistencies during integration [22]. Interoperable communication protocols and APIs are necessary for seamless data exchange and AI interaction within the metaverse. Additionally, future research should prioritise scalable and efficient algorithms for semantic data processing, enabling integration and interaction in spatiotemporal AR/VR applications for smart cities. Spatial computing can enhance interactions between IoT devices and the metaverse, but it requires ensuring data scalability and diversity. Data fusion techniques can help regenerate comprehensive views from multiple sources, improving interoperability between platforms and systems in the metaverse [284].
5.2.7. Integration with Emerging Technologies
Exploring the integration of AI with emerging technologies like 6G, quantum computing, swarm intelligence, and neuromorphic computing can enhance the capabilities of the metaverse in smart cities. These technologies can accelerate data processing and optimisation, enhance adaptability, and achieve energy-efficient operations crucial for the metaverse [37]. Implementing edge caching and federated learning techniques can minimise transmission latency and ensure data privacy. Furthermore, developing multi-agent systems (MASs) [6] and distributing AI models across end devices, edge servers, and cloud infrastructures, as shown in Figure 10, help autonomously manage dynamic changes and dependencies typical in smart cities’ rapidly evolving metaverse environments, ensuring ubiquitous experience.
5.2.8. Performance Measurement
Establishing consistent evaluation metrics and benchmarks is necessary for assessing the effectiveness of AI models in metaverse applications across various smart city domains [98]. It is also essential to develop QoE metrics tailored specifically for AR/VR devices to assess the performance of applications and services across various hardware platforms. These metrics should encompass latency, frame rate, tracking precision, and user comfort. Additionally, adopting adaptive content delivery methods can allow the metaverse to automatically modify the quality of AR/VR experiences based on the user’s device capabilities. This includes adjusting factors like resolution, detail level, and rendering quality according to the available processing power and bandwidth, ensuring optimal performance and user satisfaction [22].
5.2.9. User Education and Collaborative Governance
Enhancing user education and awareness about AI-generated content is essential to empower individuals to interact responsibly within the metaverse. Educational programmes and resources should be developed to help users critically assess and verify the authenticity of virtual content [81]. Furthermore, collaborative efforts between AI developers, policy-makers, and user communities are necessary to establish unified standards and best practices for the responsible use of generative AI technologies.
6. Conclusions
This study reveals the synergetic relationship between AI and the metaverse within the context of smart cities. By examining the fundamental domains of the metaverse, including text, visual, and audio, the profound impact of generative AI models is explored. In the visual domain, image, video, and 3D object generation emerge as key elements for creating immersive virtual experiences. The five key generative models, transformers, diffusion, autoencoders, autoregressive models, and GANs—demonstrate their ability to infuse dynamism and diversity into the metaverse environments in smart cities. Furthermore, the study delved into how these advanced AI models enable essential technologies such as big data analytics, NLP, computer vision, digital twins, IoT, Edge AI, and 5G/6G networks. These technologies collectively enhance the functionality and scalability of smart cities, fostering more efficient, sustainable, and interactive urban environments. Through the application of AI technologies in various smart city domains, real-world use cases highlight the transformative potential of an AI-enabled metaverse. Particularly, digital twins play a crucial role in promoting sustainability by enabling the accurate simulations and optimisations of urban systems and environments. However, several challenges must be addressed to fully harness the potential of an AI-enabled metaverse for smart cities. These include data acquisition, enhancing realism and ethical standards, and improving computational efficiency, privacy and interoperability. Addressing these challenges through advancements in model architectures, training algorithms, privacy/security measures, performance measures, ethical guidelines, and stakeholder collaboration is imperative for fostering responsible and advanced AI models. Ultimately, the potential of GAI models to transform the metaverse for smart cities is undeniable. Overcoming the existing challenges can unlock new dimensions of the metaverse, revolutionising digital experiences and enhancing the sustainability of smart cities worldwide.
Conceptualisation, Z.L. and J.D.; methodology, Z.L. and J.D.; validation, Z.L., J.D. and H.N.; formal analysis, Z.L.; investigation, Z.L., J.D., Q.-U.-A. and S.D.; resources, Z.L., J.D. and H.N.; data curation, Z.L. and J.D.; writing—original draft preparation, Z.L. and J.D.; writing—review and editing, Z.L., J.D., H.N., Q.-U.-A. and S.D.; visualisation, Z.L.; supervision, J.D. and H.N.; project administration, Z.L., J.D. and H.N. All authors have read and agreed to the published version of the manuscript.
No new data were created or analyzed in this study. Data sharing is not applicable.
The authors thank the Cyberspace Data and Intelligence Lab at the School of Computer and Communications Engineering, University of Science and Technology, Beijing, for providing the research facilities.
The authors declare no conflicts of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 2. Generic metaverse architecture for a smart city showing the digital, human, and physical infrastructure and the integration of the generic smart city architecture indicated in red bold square brackets.
Enlarge this image.Figure 3. Role of AI, ML, and DL techniques in the metaverse and smart city applications.
![View Image - Figure 4. Unimodal and multimodal GAI models for information creation in the metaverse [72].](https://media.proquest.com/media/hms/THUM/8/ZHa2b?_a=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&_s=5uMBeSlVDH%2FoaUHYKLf4WpKhub0%3D "View Image - Figure 4. Unimodal and multimodal GAI models for information creation in the metaverse [72].")
Enlarge this image.Figure 4. Unimodal and multimodal GAI models for information creation in the metaverse [72].
![View Image - Figure 7. A blockchain-empowered spatial crowdsourcing service in the metaverse while preserving user location privacy [185].](https://media.proquest.com/media/hms/THUM/14/ZHa2b?_a=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&_s=T2qKyeMmuXr8dOn08m%2Fvg1fSnD4%3D "View Image - Figure 7. A blockchain-empowered spatial crowdsourcing service in the metaverse while preserving user location privacy [185].")
Enlarge this image.Figure 7. A blockchain-empowered spatial crowdsourcing service in the metaverse while preserving user location privacy [185].
![View Image - Figure 8. Real-time virtual/physical synchronisation between the intelligent edge network and the metaverse [37].](https://media.proquest.com/media/hms/THUM/16/ZHa2b?_a=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&_s=CzAoYvhvClnF84Dyzogf1yHMAs8%3D "View Image - Figure 8. Real-time virtual/physical synchronisation between the intelligent edge network and the metaverse [37].")
Enlarge this image.Figure 8. Real-time virtual/physical synchronisation between the intelligent edge network and the metaverse [37].
Enlarge this image.Figure 9. Summarisation of applications for sustainable smart city in the metaverse.
Enlarge this image.Figure 10. Generic diagram of a distributed AI system for various application areas of a smart city.
Comparison of our work to other current studies concerning smart cities, AI, and the metaverse. The ✓ indicates that the review covered the corresponding techniques while ✗ indicates that review did not cover these techniques.
| Paper | Description | AI (ML/DL) | GAI/LLM | Evaluation | Big Data | NLP/CV | DT | Blockchain | IoT | Edge/5G/6G | Real-World Metaverse Smart Cities |
|---|---|---|---|---|---|---|---|---|---|---|---|
| [ | Integration of IoT-enabled technologies and AI for smart city development | ✓ | ✗ | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ |
| [ | Multi-agent-driven smart city applications | ✓ | ✗ | ✗ | ✓ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ |
| [ | Artificial intelligence for the Metaverse | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | ✗ |
| [ | Metaverse for intelligent transportation systems | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ | ✓ | ✗ |
| [ | AGI in the Metaverse for smart cities and societies | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ | ✗ | ✗ |
| [ | The Metaverse as a virtual model of platform urbanism | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ |
| [ | The potential of the Metaverse and AI for the Internet of City Things | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ |
| [ | The Metaverse as a virtual form of smart cities | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ | ✗ | ✗ |
| [ | Applications, benefits, and challenges of the Metaverse in smart cities | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✗ | ✗ | ✗ | ✗ |
| [ | Amalgamation of advanced technologies for smart city environment | ✓ | ✗ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✗ |
| [ | Metaverse applications in smart cities | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ | ✓ | ✗ |
| [ | Metaverse for smart cities | ✓ | ✗ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✗ |
| Our study | AI techniques and technologies in the Metaverse for sustainable smart city | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Distribution of research studies used in this review.
| Database Source | No. of Research Studies |
|---|---|
| IEEE Xplore | 130 |
| ACM Digital Library | 16 |
| MDPI | 24 |
| Springer | 17 |
| Science Direct | 41 |
| Other sources | 56 |
| Total | 284 |
AI-enabled technologies, techniques, and applications for metaverse-based smart cities.
| Technology | Example of AI Techniques Used | Challenges Addressed | Some Applications in Metaverse for Smart Cities |
|---|---|---|---|
| Big Data | Spatiotemporal autoregressive models, time-series clustering, GNNs, and GATs | Manages heterogeneous data from various sources, realism in virtual environments, and simulation | Real-time traffic flow prediction, environmental simulations, urban planning and infrastructure visualisation, and optimising resource allocation and customised services |
| Natural Language Processing | RNNs, CNNs, LSTM, attention mechanisms, hybrid models, and knowledge graphs | Enhances user interaction, accessibility, and personalization; machine translation; and enriches immersion through avatars | Speech-to-text and text-to-speech tasks, virtual assistants for navigation and support, chatbots, and avatars mimicking facial expressions and body language |
| Computer Vision | CNNs, GANs, and diffusion models | Rendering of avatars and scenes, object recognition/detection, and enabling VR/AR/MR | The creation of realistic avatars and 3D spaces, and enables MR experiences through holographic devices |
| Digital Twin | Multimodal models, knowledge graphs, and ML algorithms | Synchronises physical and virtual worlds, predictive maintenance, and improves immersion, simulation, and visualisation | Real-time monitoring and predictive analysis, remote operation of systems, optimising space utilisation and maintenance, and urban planning and multisystem simulations |
| Blockchain | ML techniques integrated with blockchain methods | Protects data within decentralised systems; secure data storage, sharing, and management; strengthens data integrity in digital twins and IoT; and digital economies | Managing digital assets and transactions, security in smart city infrastructures, virtual economies, secure mapping processes in IoT, and data reliability for digital twins |
| Internet of Things | Semantic communication and ML-driven semantic technologies | Real-Time Data Mapping, context awareness, interoperability, data exchange challenges, and creation of digital twins of physical elements | Real-time control of physical and virtual objects, context-aware AR/VR applications, optimises decision making, and standardising and fusing diverse urban data |
| Edge AI and 5G/6G | DL, soft computing, ML, DRL, and federated learning | Enhancing performance and efficiency at the network edge for immersive VR/AR; achieving ultra-reliable, high-speed, and low-latency communication; managing energy consumption; and enabling edge computation and IoT communication | Real-time synchronisation between physical and virtual worlds; spectrum management and utilisation efficiency; supporting mission-critical applications; immersive and interconnected metaverse experiences; optimising network performance and resource management; enhancing connectivity for autonomous vehicles; and providing pervasive intelligence with minimal latency and high bandwidth |
References
1. Akande, A.; Cabral, P.; Gomes, P.; Casteleyn, S. The Lisbon ranking for smart sustainable cities in Europe. Sustain. Cities Soc.; 2019; 44, pp. 475-487. [DOI: https://dx.doi.org/10.1016/j.scs.2018.10.009]
2. Freire, C.A.R.; Ferreira, F.A.F.; Carayannis, E.G.; Ferreira, J.J.M. Artificial Intelligence and Smart Cities: A DEMATEL Approach to Adaptation Challenges and Initiatives. IEEE Trans. Eng. Manag.; 2023; 70, pp. 1881-1899. [DOI: https://dx.doi.org/10.1109/TEM.2021.3098665]
3. Smart Cities—European Commission. Available online: https://commission.europa.eu/eu-regional-and-urban-development/topics/cities-and-urban-development/city-initiatives/smart-cities_en (accessed on 27 July 2024).
4. Singh, T.; Solanki, A.; Sharma, S.K.; Nayyar, A.; Paul, A. A Decade Review on Smart Cities: Paradigms, Challenges and Opportunities. IEEE Access; 2022; 10, pp. 68319-68364. [DOI: https://dx.doi.org/10.1109/ACCESS.2022.3184710]
5. Chiu, D.K.Y.; Xu, T.; Gondra, I. Random Graph-based Multiple Instance Learning for Structured IoT Smart City Applications. ACM Trans. Internet Technol. (TOIT); 2021; 21, 70. [DOI: https://dx.doi.org/10.1145/3448611]
6. Arora, A.; Jain, A.; Yadav, D.; Hassija, V.; Chamola, V.; Sikdar, B. Next Generation of Multi-Agent Driven Smart City Applications and Research Paradigms. IEEE Open J. Commun. Soc.; 2023; 4, pp. 2104-2121. [DOI: https://dx.doi.org/10.1109/OJCOMS.2023.3310528]
7. Mohamed, N.; Al-Jaroodi, J.; Jawhar, I.; Lazarova-Molnar, S.; Mahmoud, S. SmartCityWare: A service-oriented middleware for cloud and fog enabled smart city services. IEEE Access; 2017; 5, pp. 17576-17588. [DOI: https://dx.doi.org/10.1109/ACCESS.2017.2731382]
8. Alam, T.; Poniszewska-Maranda, A.; Maranda, W. Cloud-Based IoT Applications and Their Roles in Smart Cities. Smart Cities; 2021; 4, pp. 1196-1219. [DOI: https://dx.doi.org/10.3390/smartcities4030064]
9. Osman, A.M.S.; Elragal, A. Smart Cities and Big Data Analytics: A Data-Driven Decision-Making Use Case. Smart Cities; 2021; 4, pp. 286-313. [DOI: https://dx.doi.org/10.3390/smartcities4010018]
10. Sharif, A.; Li, J.P.; Sharif, M.I. Internet of Things network cognition and traffic management system. Cluster Comput.; 2019; 22, pp. 13209-13217. [DOI: https://dx.doi.org/10.1007/s10586-018-1722-1]
11. Bittencourt, J.C.N.; Costa, D.G.; Portugal, P.; Vasques, F. A Survey on Adaptive Smart Urban Systems. IEEE Access; 2024; 12, pp. 102826-102850. [DOI: https://dx.doi.org/10.1109/ACCESS.2024.3433381]
12. Lilhore, U.K.; Imoize, A.L.; Li, C.-T.; Simaiya, S.; Pani, S.K.; Goyal, N.; Kumar, A.; Lee, C.-C. Design and Implementation of an ML and IoT Based Adaptive Traffic-Management System for Smart Cities. Sensors; 2022; 22, 2908. [DOI: https://dx.doi.org/10.3390/s22082908]
13. Gagliardi, G.; Lupia, M.; Cario, G.; Tedesco, F.; Gaccio, F.C.; Scudo, F.L.; Casavola, A. Advanced Adaptive Street Lighting Systems for Smart Cities. Smart Cities; 2020; 3, pp. 1495-1512. [DOI: https://dx.doi.org/10.3390/smartcities3040071]
14. John, J.; Varkey, M.S.; Podder, R.S.; Sensarma, N.; Selvi, M.; Kumar, S.V.N.S.; Kannan, A. Smart Prediction and Monitoring of Waste Disposal System Using IoT and Cloud for IoT Based Smart Cities. Wirel. Pers. Commun.; 2021; 122, pp. 243-275. [DOI: https://dx.doi.org/10.1007/s11277-021-08897-z]
15. Jiang, M. An Integrated Situational Awareness Platform for Disaster Planning and Emergency Response. Proceedings of the 2020 IEEE International Smart Cities Conference, ISC2 2020; Piscataway, NJ, USA, 28 September 2020; [DOI: https://dx.doi.org/10.1109/ISC251055.2020.9239037]
16. Du, R.; Santi, P.; Xiao, M.; Vasilakos, A.V.; Fischione, C. The Sensable City: A Survey on the Deployment and Management for Smart City Monitoring. IEEE Commun. Surv. Tutor.; 2019; 21, pp. 1533-1560. [DOI: https://dx.doi.org/10.1109/COMST.2018.2881008]
17. Stephenson, N. Snow Crash: A Novel; Bantam Books: New York, NY, USA, 1992; Available online: https://search.worldcat.org/title/232304957 (accessed on 11 November 2024).
18. Wang, Y.; Su, Z.; Zhang, N.; Xing, R.; Liu, D.; Luan, T.H.; Shen, X. A Survey on Metaverse: Fundamentals, Security, and Privacy. IEEE Commun. Surv. Tutor.; 2023; 25, pp. 319-352. [DOI: https://dx.doi.org/10.1109/COMST.2022.3202047]
19. Sarwatt, D.S.; Lin, Y.; Ding, J.; Sun, Y.; Ning, H. Metaverse for Intelligent Transportation Systems (ITS): A Comprehensive Review of Technologies, Applications, Implications, Challenges and Future Directions. IEEE Trans. Intell. Transp. Syst.; 2024; 27, pp. 6290-6308. [DOI: https://dx.doi.org/10.1109/TITS.2023.3347280]
20. Wang, H.; Ning, H.; Lin, Y.; Wang, W.; Dhelim, S.; Farha, F.; Ding, J.; Daneshmand, M. A Survey on the Metaverse: The State-of-the-Art, Technologies, Applications, and Challenges. IEEE Internet Things J.; 2023; 10, pp. 14671-14688. [DOI: https://dx.doi.org/10.1109/JIOT.2023.3278329]
21. Milgram, P.; Takemura, H.; Utsumi, A.; Kishino, F. Augmented reality: A class of displays on the reality-virtuality continuum. Telemanipulator and Telepresence Technologies; SPIE: Philadelphia, PA, USA, 1995; Volume 2351, pp. 282-292. [DOI: https://dx.doi.org/10.1117/12.197321]
22. Li, K.; Lau, B.P.L.; Yuan, X.; Ni, W.; Guizani, M.; Yuen, C. Toward Ubiquitous Semantic Metaverse: Challenges, Approaches, and Opportunities. IEEE Internet Things J.; 2023; 10, pp. 21855-21872. [DOI: https://dx.doi.org/10.1109/JIOT.2023.3302159]
23. Ng, W.C.; Du, H.; Lim, W.Y.B.; Xiong, Z.; Niyato, D.; Miao, C. Stochastic Resource Allocation for Semantic Communication-Aided Virtual Transportation Networks in the Metaverse. Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC); Dubai, United Arab Emirates, 21–24 April 2024; [DOI: https://dx.doi.org/10.1109/WCNC57260.2024.10570867]
24. Kountouris, M.; Pappas, N. Semantics-Empowered Communication for Networked Intelligent Systems. IEEE Commun. Mag.; 2021; 59, pp. 96-102. [DOI: https://dx.doi.org/10.1109/MCOM.001.2000604]
25. Manolova, A.; Tonchev, K.; Poulkov, V.; Dixir, S.; Lindgren, P. Context-Aware Holographic Communication Based on Semantic Knowledge Extraction. Wirel. Pers. Commun.; 2021; 120, pp. 2307-2319. [DOI: https://dx.doi.org/10.1007/s11277-021-08560-7]
26. Maatouk, A.; Assaad, M.; Ephremides, A. The Age of Incorrect Information: An Enabler of Semantics-Empowered Communication. IEEE Trans. Wirel. Commun.; 2023; 22, pp. 2621-2635. [DOI: https://dx.doi.org/10.1109/TWC.2022.3213227]
27. Wang, J.; Du, H.; Tian, Z.; Niyato, D.; Kang, J.; Shen, X. Semantic-Aware Sensing Information Transmission for Metaverse: A Contest Theoretic Approach. IEEE Trans. Wirel. Commun.; 2023; 22, pp. 5214-5228. [DOI: https://dx.doi.org/10.1109/TWC.2022.3232565]
28. Chang, L.; Zhang, Z.; Li, P.; Xi, S.; Guo, W.; Shen, Y.; Xiong, Z.; Kang, J.; Niyato, D.; Qiao, X. et al. 6G-Enabled Edge AI for Metaverse: Challenges, Methods, and Future Research Directions. J. Commun. Inf. Netw.; 2022; 7, pp. 107-121. [DOI: https://dx.doi.org/10.23919/JCIN.2022.9815195]
29. Allam, Z.; Sharifi, A.; Bibri, S.E.; Jones, D.S.; Krogstie, J. The Metaverse as a Virtual Form of Smart Cities: Opportunities and Challenges for Environmental, Economic, and Social Sustainability in Urban Futures. Smart Cities; 2022; 5, pp. 771-801. [DOI: https://dx.doi.org/10.3390/smartcities5030040]
30. Zawish, M.; Dharejo, F.A.; Khowaja, S.A.; Raza, S.; Davy, S.; Dev, K.; Bellavista, P. AI and 6G Into the Metaverse: Fundamentals, Challenges and Future Research Trends. IEEE Open J. Commun. Soc.; 2024; 5, pp. 730-778. [DOI: https://dx.doi.org/10.1109/OJCOMS.2024.3349465]
31. Li, K.; Cui, Y.; Li, W.; Lv, T.; Yuan, X.; Li, S.; Ni, W.; Simsek, M.; Dressler, F. When Internet of Things Meets Metaverse: Convergence of Physical and Cyber Worlds. IEEE Internet Things J.; 2023; 10, pp. 4148-4173. [DOI: https://dx.doi.org/10.1109/JIOT.2022.3232845]
32. Wang, X.; Huang, J.; Tian, Y.L.; Wang, F.Y. AGI in Metaverse for Smart Cities and Societies: A Cyber Physical Social Approach. Proceedings of the 2024 Australian and New Zealand Control Conference; Gold Coast, Australia, 1–2 February 2024; pp. 61-66. [DOI: https://dx.doi.org/10.1109/ANZCC59813.2024.10432879]
33. Bibri, S.E. The Metaverse as a Virtual Model of Platform Urbanism: Its Converging AIoT, XReality, Neurotech, and Nanobiotech and Their Applications, Challenges, and Risks. Smart Cities; 2023; 6, pp. 1345-1384. [DOI: https://dx.doi.org/10.3390/smartcities6030065]
34. Bibri, S.E.; Jagatheesaperumal, S.K. Harnessing the Potential of the Metaverse and Artificial Intelligence for the Internet of City Things: Cost-Effective XReality and Synergistic AIoT Technologies. Smart Cities; 2023; 6, pp. 2397-2429. [DOI: https://dx.doi.org/10.3390/smartcities6050109]
35. Maier, F.; Weinberger, M. Metaverse Meets Smart Cities—Applications, Benefits, and Challenges. Future Internet; 2024; 16, 126. [DOI: https://dx.doi.org/10.3390/fi16040126]
36. Rani, S.; Mishra, R.K.; Usman, M.; Kataria, A.; Kumar, P.; Bhambri, P.; Mishra, A.K. Amalgamation of advanced technologies for sustainable development of smart city environment: A review. IEEE Access; 2021; 9, pp. 150060-150087. [DOI: https://dx.doi.org/10.1109/ACCESS.2021.3125527]
37. Ali, M.; Naeem, F.; Kaddoum, G.; Hossain, E. Metaverse Communications, Networking, Security, and Applications: Research Issues, State-of-the-Art, and Future Directions. IEEE Commun. Surv. Tutor.; 2024; 26, pp. 1238-1278. [DOI: https://dx.doi.org/10.1109/COMST.2023.3347172]
38. Yaqoob, I.; Salah, K.; Jayaraman, R.; Omar, M. Metaverse applications in smart cities: Enabling technologies, opportunities, challenges, and future directions. Internet Things; 2023; 23, 100884. [DOI: https://dx.doi.org/10.1016/j.iot.2023.100884]
39. Alahi, M.E.E.; Sukkuea, A.; Tina, F.W.; Nag, A.; Kurdthongmee, W.; Suwannarat, K.; Mukhopadhyay, S.C. Integration of IoT-Enabled Technologies and Artificial Intelligence (AI) for Smart City Scenario: Recent Advancements and Future Trends. Sensors; 2023; 23, 5206. [DOI: https://dx.doi.org/10.3390/s23115206]
40. Huynh-The, T.; Pham, Q.V.; Pham, X.Q.; Nguyen, T.T.; Han, Z.; Kim, D.S. Artificial intelligence for the metaverse: A survey. Eng. Appl. Artif. Intell.; 2023; 117, 105581. [DOI: https://dx.doi.org/10.1016/j.engappai.2022.105581]
41. Chen, Z.; Gan, W.; Wu, J.; Lin, H.; Chen, C.M. Metaverse for smart cities: A survey. Internet Things Cyber-Phys. Syst.; 2024; 4, pp. 203-216. [DOI: https://dx.doi.org/10.1016/j.iotcps.2023.12.002]
42. Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E. et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ; 2021; 372, 105918. [DOI: https://dx.doi.org/10.1136/BMJ.N71]
43. Kozhevnikov, S.; Svitek, M. From Smart City sustainable development to resiliency by-design. Proceedings of the 2022 Smart Cities Symposium Prague (SCSP); Prague, Czech Republic, 26–27 May 2022; [DOI: https://dx.doi.org/10.1109/SCSP54748.2022.9792566]
44. Ammara, U.; Rasheed, K.; Mansoor, A.; Al-Fuqaha, A.; Qadir, J. Smart Cities from the Perspective of Systems. Systems; 2022; 10, 77. [DOI: https://dx.doi.org/10.3390/systems10030077]
45. Svítek, M.; Skobelev, P.; Kozhevnikov, S. Smart City 5.0 as an Urban Ecosystem of Smart Services. Stud. Comput. Intell.; 2020; 853, pp. 426-438. [DOI: https://dx.doi.org/10.1007/978-3-030-27477-1_33]
46. Khan, A.; Aslam, S.; Aurangzeb, K.; Alhussein, M.; Javaid, N. Multiscale modeling in smart cities: A survey on applications, current trends, and challenges. Sustain. Cities Soc.; 2022; 78, 103517. [DOI: https://dx.doi.org/10.1016/j.scs.2021.103517]
47. Al Sharif, R.; Pokharel, S. Smart City Dimensions and Associated Risks: Review of literature. Sustain. Cities Soc.; 2022; 77, 103542. [DOI: https://dx.doi.org/10.1016/j.scs.2021.103542]
48. Chen, S.; Wei, Z.; Sun, G.; Cheung, K.W.; Wang, D.; Zang, H. Adaptive Robust Day-Ahead Dispatch for Urban Energy Systems. IEEE Trans. Ind. Electron.; 2019; 66, pp. 1379-1390. [DOI: https://dx.doi.org/10.1109/TIE.2017.2787605]
49. De Carolis, B.; Ladogana, F.; MacChiarulo, N. YOLO TrashNet: Garbage Detection in Video Streams. Proceedings of the IEEE Conference on Evolving and Adaptive Intelligent Systems; Bari, Italy, 27–29 May 2020; [DOI: https://dx.doi.org/10.1109/EAIS48028.2020.9122693]
50. Silva, B.N.; Khan, M.; Han, K. Towards sustainable smart cities: A review of trends, architectures, components, and open challenges in Smart Cities. Sustain. Cities Soc.; 2018; 38, pp. 697-713. [DOI: https://dx.doi.org/10.1016/j.scs.2018.01.053]
51. Yang, F.; Hua, Y.; Li, X.; Yang, Z.; Yu, X.; Fei, T. A survey on multisource heterogeneous urban sensor access and data management technologies. Meas. Sens.; 2022; 19, 100061. [DOI: https://dx.doi.org/10.1016/j.measen.2021.100061]
52. Da Silva, G.F.P.; Costa, D.G.; De Jesus, T.C. A Secure OTA Approach for Flexible Operation of Emergency Detection Units in Smart Cities. Proceedings of the 2023 IEEE International Smart Cities Conference (ISC2); Bucharest, Romania, 24–27 September 2023; [DOI: https://dx.doi.org/10.1109/ISC257844.2023.10293637]
53. Wang, S.; Cao, J.; Yu, P.S. Deep Learning for Spatio-Temporal Data Mining: A Survey. IEEE Trans. Knowl. Data Eng.; 2022; 34, pp. 3681-3700. [DOI: https://dx.doi.org/10.1109/TKDE.2020.3025580]
54. The Metaverse Value-Chain. Trillions of Dollars are Pouring Into… | by Jon Radoff | Building the Metaverse | Medium. Available online: https://medium.com/building-the-metaverse/the-metaverse-value-chain-afcf9e09e3a7 (accessed on 12 August 2024).
55. Genay, A.; Lecuyer, A.; Hachet, M. Being an Avatar ‘for Real’: A Survey on Virtual Embodiment in Augmented Reality. IEEE Trans. Vis. Comput. Graph.; 2022; 28, pp. 5071-5090. [DOI: https://dx.doi.org/10.1109/TVCG.2021.3099290]
56. Kai, C.; Zhou, H.; Yi, Y.; Huang, W. Collaborative Cloud-Edge-End Task Offloading in Mobile-Edge Computing Networks with Limited Communication Capability. IEEE Trans. Cogn. Commun. Netw.; 2021; 7, pp. 624-634. [DOI: https://dx.doi.org/10.1109/TCCN.2020.3018159]
57. IEEE 2888—Home. Available online: https://sagroups.ieee.org/2888/ (accessed on 8 September 2024).
58. Baccour, E.; Mhaisen, N.; Abdellatif, A.A.; Erbad, A.; Mohamed, A.; Hamdi, M.; Guizani, M. Pervasive AI for IoT Applications: A Survey on Resource-Efficient Distributed Artificial Intelligence. IEEE Commun. Surv. Tutor.; 2022; 24, pp. 2366-2418. [DOI: https://dx.doi.org/10.1109/COMST.2022.3200740]
59. Russell, S.; Norvig, P. Artificial Intelligence: A Modern Approach; 4th ed. Pearson: Madrid, Spain, 2021; Available online: https://aima.cs.berkeley.edu/ (accessed on 22 November 2024).
60. Yu, Z.; Abdulghani, A.M.; Zahid, A.; Heidari, H.; Imran, M.A.; Abbasi, Q.H. An overview of neuromorphic computing for artificial intelligence enabled hardware-based hopfield neural network. IEEE Access; 2020; 8, pp. 67085-67099. [DOI: https://dx.doi.org/10.1109/ACCESS.2020.2985839]
61. Lecun, Y.; Bengio, Y. Deep learning. Nature; 2015; 521, pp. 436-444. [DOI: https://dx.doi.org/10.1038/nature14539]
62. Shrestha, A.; Mahmood, A. Review of deep learning algorithms and architectures. IEEE Access; 2019; 7, pp. 53040-53065. [DOI: https://dx.doi.org/10.1109/ACCESS.2019.2912200]
63. LeCun, Y.; Bengio, Y. Convolutional networks for images, speech, and time series. The Handbook of Brain Theory and Neural Networks; MIT Press: Cambridge, MA, USA, 1995; Volume 3361, 1995.
64. Medsker, L.; Jain, L. Recurrent neural networks. Des. Appl.; 2001; 5, pp. 64-67.
65. Socher, R.; Perelygin, A.; Wu, J.; Chuang, J. Recursive deep models for semantic compositionality over a sentiment treebank. Proceedings of the 2013 Conference on Empirical Methods in Natural; Washington, DC, USA, 28 March 2013; Volume 5, pp. 1631-1642. Available online: https://aclanthology.org/D13-1170.pdf (accessed on 20 November 2024).
66. Polydoros, A.S.; Nalpantidis, L. Survey of Model-Based Reinforcement Learning: Applications on Robotics. J. Intell. Robot. Syst. Theory Appl.; 2017; 86, pp. 153-173. [DOI: https://dx.doi.org/10.1007/s10846-017-0468-y]
67. Baevski, A.; Hsu, W.N.; Xu, Q.; Babu, A.; Gu, J.; Auli, M. Data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language. Proc. Mach. Learn. Res.; 2022; 162, pp. 1298-1312.
68. Yang, L.; Luo, Z.; Zhang, S.; Teng, F.; Li, T. Continual Learning for Smart City: A Survey. arXiv; 2024; arXiv: 2404.00983[DOI: https://dx.doi.org/10.1109/TKDE.2024.3447123]
69. Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A.N.; Kaiser, L.; Polosukhin, I. Attention is All you Need. Proceedings of the Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems; Long Beach, CA, USA, 12 June 2017.
70. Feng, C.; Chen, H.; Cai, F.; De Rijke, M. Attentive encoder-based extractive text summarization. Proceedings of the International Conference on Information and Knowledge Management; Turin, Italy, 22–26 October 2018; pp. 1499-1502. [DOI: https://dx.doi.org/10.1145/3269206.3269251]
71. Raffel, C.; Shazeer, N.; Roberts, A.; Lee, K.; Narang, S.; Matena, M.; Zhou, Y.; Li, W.; Liu, P.J. Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res.; 2020; 21, pp. 1-67. Available online: https://dl.acm.org/doi/abs/10.5555/3455716.3455856 (accessed on 23 March 2023).
72. Liang, C.; Du, H.; Sun, Y.; Niyato, D.; Kang, J.; Zhao, D.; Imran, M.A. Generative AI-driven Semantic Communication Networks: Architecture, Technologies and Applications. IEEE Trans. Cogn. Commun. Netw.; 2024; [DOI: https://dx.doi.org/10.1109/TCCN.2024.3435524]
73. Hagos, D.H.; Battle, R.; Rawat, D.B. Recent Advances in Generative AI and Large Language Models: Current Status, Challenges, and Perspectives. IEEE Trans. Artif. Intell.; 2024; pp. 1-21. [DOI: https://dx.doi.org/10.1109/TAI.2024.3444742]
74. Liu, Y.; Han, T.; Ma, S.; Zhang, J.; Yang, Y.; Tian, J.; He, H.; Li, A.; He, M.; Liu, Z. et al. Summary of ChatGPT-Related research and perspective towards the future of large language models. Meta-Radiol.; 2023; 1, 100017. [DOI: https://dx.doi.org/10.1016/j.metrad.2023.100017]
75. Du, H.; Teng, S.; Chen, H.; Ma, J.; Wang, X.; Gou, C.; Li, B.; Ma, S.; Miao, Q.; Na, X. et al. Chat With ChatGPT on Intelligent Vehicles: An IEEE TIV Perspective. IEEE Trans. Intell. Veh.; 2023; 8, pp. 2020-2026. [DOI: https://dx.doi.org/10.1109/TIV.2023.3253281]
76. Chang, Y.; Wang, X.; Wang, J.; Wu, Y.; Yang, L.; Zhu, K.; Chen, H.; Yi, X.; Wang, C.; Wang, Y. et al. A Survey on Evaluation of Large Language Models. ACM Trans. Intell. Syst. Technol.; 2023; 15, 39. [DOI: https://dx.doi.org/10.1145/3641289]
77. Min, B.; Ross, H.; Sulem, E.; Ben Veyseh, A.P.; Nguyen, T.H.; Sainz, O.; Agirre, E.; Heintz, I.; Roth, D. Recent Advances in Natural Language Processing via Large Pre-trained Language Models: A Survey. ACM Comput. Surv.; 2023; 56, 1145. [DOI: https://dx.doi.org/10.1145/3605943]
78. Thirunavukarasu, A.J.; Ting, D.S.J.; Elangovan, K.; Gutierrez, L.; Tan, T.F.; Ting, D.S.W. Large language models in medicine. Nat. Med.; 2023; 29, pp. 1930-1940. [DOI: https://dx.doi.org/10.1038/s41591-023-02448-8] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37460753]
79. Wang, S.; Zhu, Y.; Li, Z.; Wang, Y.; Li, L.; He, Z. ChatGPT as Your Vehicle Co-Pilot: An Initial Attempt. IEEE Trans. Intell. Veh.; 2023; 8, pp. 4706-4721. [DOI: https://dx.doi.org/10.1109/TIV.2023.3325300]
80. Tian, Y.; Li, X.; Zhang, H.; Zhao, C.; Li, B.; Wang, X.; Wang, F.-Y. VistaGPT: Generative Parallel Transformers for Vehicles With Intelligent Systems for Transport Automation. IEEE Trans. Intell. Veh.; 2023; 8, pp. 4198-4207. [DOI: https://dx.doi.org/10.1109/TIV.2023.3307012]
81. Chamola, V.; Bansal, G.; Das, T.K.; Hassija, V.; Sai, S.; Wang, J.; Zeadally, S.; Hussain, A.; Yu, F.R.; Guizani, M. et al. Beyond Reality: The Pivotal Role of Generative AI in the Metaverse. IEEE Internet Things Mag.; 2024; 7, pp. 126-135. [DOI: https://dx.doi.org/10.1109/IOTM.001.2300174]
82. Suzuki, M.; Matsuo, Y. A survey of multimodal deep generative models. Adv. Robot.; 2022; 36, pp. 261-278. [DOI: https://dx.doi.org/10.1080/01691864.2022.2035253]
83. Baltrusaitis, T.; Ahuja, C.; Morency, L.P. Multimodal Machine Learning: A Survey and Taxonomy. IEEE Trans. Pattern Anal. Mach. Intell.; 2019; 41, pp. 423-443. [DOI: https://dx.doi.org/10.1109/TPAMI.2018.2798607]
84. Radford, A.; Wu, J.; Child, R.; Luan, D.; Amodei, D.; Sutskever, I. Language Models are Unsupervised Multitask Learners. Available online: https://github.com/codelucas/newspaper (accessed on 26 April 2023).
85. Brown, T.B. Language Models are Few-Shot Learners. Proceedings of the Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems; Virtual, 6–12 December 2020; pp. 1877-1901. Available online: https://commoncrawl.org/the-data/ (accessed on 29 October 2023).
86. Devlin, J.; Chang, M.-W.; Lee, K.; Google, K.T.; Language, A.I. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Proceedings of the 2019 Conference of the North; Minnesota, MN, USA, 2–7 June 2019; pp. 4171-4186. [DOI: https://dx.doi.org/10.18653/V1/N19-1423]
87. Bahdanau, D.; Cho, K.H.; Bengio, Y. Neural Machine Translation by Jointly Learning to Align and Translate. arXiv; 2015; arXiv: 1409.0473
88. Sutskever, I.; Vinyals, O.; Le, Q.V. Sequence to sequence learning with neural networks. Proceedings of the Advances in Neural Information Processing Systems 4: Annual Conference on Neural Information Processing Systems; Montreal, QC, Canada, 8–13 December 2014; pp. 3104-3112.
89. Vinyals, O.; Le, Q. A Neural Conversational Model. arXiv; 2015; arXiv: 1506.05869
90. Dai, S.; Gan, Z.; Cheng, Y.; Tao, C.; Carin, L.; Liu, J. APo-VAE: Text Generation in Hyperbolic Space. Proceedings of the NAACL-HLT 2021—2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies; Online, 6–11 June 2021; pp. 416-431. [DOI: https://dx.doi.org/10.18653/v1/2021.naacl-main.36]
91. Zhang, X.; Yang, Y.; Yuan, S.; Shen, D.; Carin, L. Syntax-Infused Variational Autoencoder for Text Generation. Proceedings of the ACL 2019—57th Annual Meeting of the Association for Computational Linguistics; Florence, Italy, 28 July–2 August 2019; pp. 2069-2078. [DOI: https://dx.doi.org/10.18653/v1/p19-1199]
92. Wang, W.; Gan, Z.; Xu, H.; Zhang, R.; Wang, G.; Shen, D.; Chen, C.; Carin, L. Topic-Guided Variational Autoencoders for Text Generation. Proceedings of the NAACL HLT 2019—2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies; Minneapolis, MN, USA, 2–7 June 2019; Volume 1, pp. 166-177. [DOI: https://dx.doi.org/10.18653/v1/n19-1015]
93. Kingma, D.P.; Salimans, T.; Jozefowicz, R.; Chen, X.; Sutskever, I.; Welling, M. Improved Variational Inference with Inverse Autoregressive Flow. Proceedings of the Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems; Barcelona, Spain, 5–10 December 2016.
94. Guo, J.; Lu, S.; Cai, H.; Zhang, W.; Yu, Y.; Wang, J. Long Text Generation via Adversarial Training with Leaked Information. Proceedings of the AAAI Conference on Artificial Intelligence; New Orleans, LA, USA, 2–7 February 2018; Volume 32, pp. 5141-5148. [DOI: https://dx.doi.org/10.1609/AAAI.V32I1.11957]
95. Wang, H.; Qin, Z.; Wan, T. Text Generation Based on Generative Adversarial Nets with Latent Variables. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); LNAI: Burlingame, CA, USA, 2018; Volume 10938, pp. 92-103. [DOI: https://dx.doi.org/10.1007/978-3-319-93037-4_8]
96. Creswell, A.; White, T.; Dumoulin, V.; Arulkumaran, K.; Sengupta, B.; Bharath, A.A. Generative Adversarial Networks: An Overview. IEEE Signal Process. Mag.; 2018; 35, pp. 53-65. [DOI: https://dx.doi.org/10.1109/MSP.2017.2765202]
97. Goodfellow, I.; Pouget-Abadie, J.; Mirza, M.; Xu, B.; Warde-Farley, D.; Ozair, S.; Courville, A.; Bengio, Y. Generative adversarial nets. Proceedings of the Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Processing Systems; Montreal, QC, Canada, 8–13 December 2014; Available online: https://proceedings.neurips.cc/paper/5423-generative-adversarial-nets (accessed on 20 November 2024).
98. Bandi, A.; Adapa, P.V.S.R.; Kuchi, Y.E.V.P.K. The Power of Generative AI: A Review of Requirements, Models, Input–Output Formats, Evaluation Metrics, and Challenges. Future Internet; 2023; 15, 260. [DOI: https://dx.doi.org/10.3390/fi15080260]
99. Song, J.; Meng, C.; Ermon, S. Denoising Diffusion Implicit Models. arXiv; 2020; arXiv: 2010.02502
100. Ho, J.; Jain, A. Denoising diffusion probabilistic models. Proceedings of the Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems; Virtual, 6–12 December 2020; pp. 6840-6851. Available online: https://proceedings.neurips.cc/paper/2020/hash/4c5bcfec8584af0d967f1ab10179ca4b-Abstract.html (accessed on 19 November 2024).
101. Zach, C. Fully Variational Noise-Contrastive Estimation. Image Analysis; Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) LNCS Springer: Cham, Switzerland, 2023; 13886, pp. 175-190. [DOI: https://dx.doi.org/10.1007/978-3-031-31438-4_12]
102. Song, Y.; Sohl-Dickstein, J.; Kingma, D.P.; Kumar, A.; Ermon, S.; Poole, B. Score-Based Generative Modeling through Stochastic Differential Equations. arXiv; 2020; arXiv: 2011.13456
103. Gong, S.; Li, M.; Feng, J.; Wu, Z.; Kong, L. Diffuseq: Sequence to Sequence Text Generation with Diffusion Models. Proceedings of the 11th International Conference on Learning Representations (ICLR); Kigali, Rwanda, 1–5 May 2023.
104. Li, X.; Thickstun, J.; Gulrajani, I.; Liang, P.S.; Hashimoto, T.B. Diffusion-LM Improves Controllable Text Generation. Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems; New Orleans, LA, USA, 28 November–9 December 2022; pp. 4328-4343.
105. Chen, J.; Guo, H.; Yi, K.; Li, B.; Elhoseiny, M. VisualGPT: Data-Efficient Adaptation of Pretrained Language Models for Image Captioning. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; New Orleans, LA, USA, 19–21 June 2022.
106. Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S. et al. An Image is Worth 16×16 Words: Transformers for Image Recognition at Scale. Proceedings of the ICLR 2021—9th International Conference on Learning Representations; Virtual Event, Austria, 3–7 May 2021.
107. Xu, H.; Ghosh, G.; Huang, P.-Y.; Okhonko, D.; Aghajanyan, A.; Metze, F.; Zettlemoyer, L.; Feichtenhofer, C. VideoCLIP: Contrastive Pre-training for Zero-shot Video-Text Understanding. Proceedings of the EMNLP 2021—2021 Conference on Empirical Methods in Natural Language Processing; Online, 7–11 November 2021; pp. 6787-6800. [DOI: https://dx.doi.org/10.18653/v1/2021.emnlp-main.544]
108. Hannun, A. Deep Speech: Scaling Up End-to-End Speech Recognition. arXiv; 2014; arXiv: 1412.5567
109. Baevski, A.; Zhou, Y.; Mohamed, A.; Auli, M. Wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations. Proceedings of the Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems; Virtual, 6–12 December 2020; pp. 12449-12460.
110. van den Oord, A.; Kalchbrenner, N.; Kavukcuoglu, K. Pixel Recurrent Neural Networks. PMLR; 2016; 48, pp. 1747-1756.
111. Van den Oord, A.; Kalchbrenner, N.; Espeholt, L.; Vinyals, O.; Graves, A. Conditional image generation with pixelcnn decoders. Proceedings of the Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems; Barcelona, Spain, 5–10 December 2016; pp. 4797-4805.
112. Huang, H.; He, R.; Sun, Z.; Tan, T. Introvae: Introspective variational autoencoders for photographic image synthesis. Proceedings of the Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems; Montréal, QC, Canada, 3–8 December 2018; pp. 52-63.
113. Razavi, A.; Van den Oord, A.; Vinyals, O. Generating diverse high-fidelity images with vq-vae-2. Proceedings of the Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems; Vancouver, BC, Canada, 8–14 December 2019; pp. 14866-14876.
114. Zhu, J.-Y.; Park, T.; Isola, P.; Efros, A.A. Unpaired Image-To-Image Translation Using Cycle-Consistent Adversarial Networks. Proceedings of the IEEE International Conference on Computer Vision; Venice, Italy, 22–29 October 2017; pp. 2223-2232.
115. Karras, T.; Laine, S.; Aila, T. A Style-Based Generator Architecture for Generative Adversarial Networks. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; Long Beach, CA, USA, 15–20 June 2019; Volume 43, pp. 4401-4410.
116. Bao, J.; Chen, D.; Wen, F.; Li, H.; Hua, G. CVAE-GAN: Fine-Grained Image Generation Through Asymmetric Training. Proceedings of the IEEE International Conference on Computer Vision; Venice, Italy, 22–29 October 2017; pp. 2745-2754.
117. Kingma, D.P.; Dhariwal, P. Glow: Generative Flow with Invertible 1x1 Convolutions. Proceedings of the Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems; Montréal, QC, Canada, 3–8 December 2018.
118. Dinh, L.; Sohl-Dickstein, J.; Bengio, S. Density estimation using Real NVP. arXiv; 2016; arXiv: 1605.08803
119. Ramesh, A.; Pavlov, M.; Goh, G.; Gray, S.; Voss, C.; Radford, A.; Chen, M.; Sutskever, I. Zero-Shot Text-to-Image Generation. PMLR; 2021; 139, pp. 8821-8831.
120. Ramesh, A.; Dhariwal, P.; Nichol, A.; Chu, C.; Chen, M. Hierarchical Text-Conditional Image Generation with CLIP Latents. arXiv; 2022; arXiv: 2204.06125
121. Radford, A.; Kim, J.W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J. et al. Learning Transferable Visual Models from Natural Language Supervision. Proceedings of the International conference on machine learning; Virtual, 18–24 July 2021.
122. Ho, J.; Salimans, T.; Gritsenko, A.; Chan, W.; Norouzi, M.; Fleet, D.J. Video Diffusion Models. Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems; New Orleans, LA, USA, 28 November–9 December 2022; pp. 8633-8646.
123. Tulyakov, S.; Liu, M.-Y.; Yang, X.; Kautz, J. MoCoGAN: Decomposing Motion and Content for Video Generation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; Salt Lake City, Utah, 18–22 June 2018; pp. 1526-1535.
124. Villegas, R.; Babaeizadeh, M.; Kindermans, P.J.; Moraldo, H.; Zhang, H.; Saffar, M.T.; Castro, S.; Kunze, J.; Erhan, D. Phenaki: Variable Length Video Generation from Open Domain Textual Descriptions. Proceedings of the International Conference on Learning Representations; Virtual, 25 April 2022.
125. Brooks, T.; Peebles, B.; Holmes, C.; DePue, W.; Guo, Y.; Jing, L.; Schnurr, D.; Taylor, J.; Luhman, T.; Luhman, E. et al. Video Generation Models as World Simulators. Available online: https://openai.com/index/video-generation-models-as-world-simulators/ (accessed on 20 November 2024).
126. Hong, W.; Ding, M.; Zheng, W.; Liu, X.; Tang, J. CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers. arXiv; 2022; arXiv: 2205.15868
127. Wu, J.Z.; Ge, Y.; Wang, X.; Lei, S.W.; Gu, Y.; Shi, Y.; Hsu, W.; Shan, Y.; Qie, X.; Shou, M.Z. Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation. Proceedings of the IEEE/CVF International Conference on Computer Vision; Paris, France, 2–6 October 2023.
128. Karras, T.; Laine, S.; Aittala, M.; Hellsten, J.; Lehtinen, J.; Aila, T. Analyzing and Improving the Image Quality of StyleGAN. 2020; Available online: https://github.com/NVlabs/stylegan2 (accessed on 14 August 2024).
129. Gal, R.; Patashnik, O.; Maron, H.; Bermano, A.H.; Chechik, G.; Cohen, D.-O. StyleGAN-NADA: CLIP-Guided Domain Adaptation of Image Generators. ACM Trans. Graph.; 2022; 41, 141. [DOI: https://dx.doi.org/10.1145/3528223.3530164]
130. Saharia, C.; Chan, W.; Saxena, S.; Li, L.; Whang, J.; Denton, E.L.; Ghasemipour, K.; Gontijo Lopes, R.; Karagol Ayan, B.; Salimans, T. et al. Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding. Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems; New Orleans, LA, USA, 28 November–9 December 2022; pp. 36479-36494.
131. Lin, C.H.; Gao, J.; Tang, L.; Takikawa, T.; Zeng, X.; Huang, X.; Kreis, K.; Fidler, S.; Liu, M.Y.; Lin, T.Y. Magic3D: High-Resolution Text-to-3D Content Creation. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; Vancouver, BC, Canada, 17–24 June 2023; pp. 300-309.
132. Poole, B.; Jain, A.; Barron, J.T.; Mildenhall, B. DreamFusion: Text-to-3D using 2D Diffusion. arXiv; 2022; arXiv: 2209.14988
133. Müller, T.; Evans, A.; Schied, C.; Keller, A. Instant neural graphics primitives with a multiresolution hash encoding. ACM Trans. Graph.; 2022; 41, 102. [DOI: https://dx.doi.org/10.1145/3528223.3530127]
134. Chen, H.; Gu, J.; Chen, A.; Tian, W.; Tu, Z.; Liu, L.; Su, H. Single-Stage Diffusion NeRF: A Unified Approach to 3D Generation and Reconstruction. Proceedings of the IEEE/CVF International Conference on Computer Vision; Paris, France, 2–6 October 2023; pp. 2416-2425.
135. Lee, J.L.; Li, C.; DiSR-NeRF, G.H.L. DiSR-NeRF: Diffusion-Guided View-Consistent Super-Resolution NeRF. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; Seattle, WA, USA, 16–22 June 2024.
136. Pan, X.; Yang, Z.; Bai, S.; Yang, Y. GD^2-NeRF: Generative Detail Compensation via GAN and Diffusion for One-shot Generalizable Neural Radiance Fields. arXiv; 2024; arXiv: 2401.00616
137. van Den Oord, A.; Dieleman, S.; Zen, H.; Simonyan, K.; Vinyals, O.; Graves, A.; Kalchbrenner, N.; Senior, A.; Kavukcuoglu, K. WaveNet: A Generative Model for Raw Audio. September 2016. Available online: http://arxiv.org/abs/1609.03499 (accessed on 19 November 2024).
138. Mehri, S.; Kumar, K.; Gulrajani, I.; Kumar, R.; Jain, S.; Sotelo, J.; Courville, A.; Bengio, Y. SampleRNN: An Unconditional End-to-End Neural Audio Generation Model. arXiv; 2016; arXiv: 1612.07837
139. Engel, J.; Agrawal, K.K.; Chen, S.; Gulrajani, I.; Donahue, C.; Roberts, A. GANSynth: Adversarial Neural Audio Synthesis. Proceedings of the 7th International Conference on Learning Representations (ICLR); New Orleans, LA, USA, 6–9 May 2019.
140. Donahue, C.; McAuley, J.; Puckette, M. Adversarial Audio Synthesis. Proceedings of the 7th International Conference on Learning Representations (ICLR 2019); New Orleans, LA, USA, 6–9 May 2019.
141. Schneider, F. ArchiSound: Audio Generation with Diffusion. arXiv; 2023; arXiv: 2301.13267
142. Meng, Y.; Li, W.; Lei, S.; Zou, Z.; Shi, Z. Large-Factor Super-Resolution of Remote Sensing Images with Spectra-Guided Generative Adversarial Networks. IEEE Trans. Geosci. Remote Sens.; 2022; 60, 5634111. [DOI: https://dx.doi.org/10.1109/TGRS.2022.3222360]
143. Akuzawa, K.; Onishi, K. Conditional deep hierarchical variational autoencoder for voice conversion. Proceedings of the 2021 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC); Tokyo, Japan, 14–17 December 2021; pp. 808-813.
144. Borsos, Z.; Marinier, R.; Vincent, D.; Kharitonov, E.; Pietquin, O.; Sharifi, M.; Roblek, D.; Teboul, O.; Grangier, D.; Tagliasacchi, M. et al. AudioLM: A Language Modeling Approach to Audio Generation. IEEE/ACM Trans. Audio Speech Lang. Process.; 2023; 31, pp. 2523-2533. [DOI: https://dx.doi.org/10.1109/TASLP.2023.3288409]
145. Wang, Y.; Skerry-Ryan, R.J.; Stanton, D.; Wu, Y.; Weiss, R.J.; Jaitly, N.; Yang, Z.; Xiao, Y.; Chen, Z.; Bengio, S. et al. Tacotron: Towards End-to-End Speech Synthesis. Proceedings of the Annual Conference of the International Speech Communication Association; Stockholm, Sweden, 20–24 August 2017; pp. 4006-4010. [DOI: https://dx.doi.org/10.21437/Interspeech.2017-1452]
146. Google Assistant, Your Own Personal Google. Available online: https://assistant.google.com/ (accessed on 19 November 2024).
147. Amazon Alexa Voice AI | Alexa Developer Official Site. Available online: https://developer.amazon.com/en-US/alexa (accessed on 19 November 2024).
148. Siri—Apple. Available online: https://www.apple.com/siri/ (accessed on 19 November 2024).
149. DuerOS Overview—DuerOS. Available online: https://dueros.baidu.com/en/html/dueros/index.html (accessed on 19 November 2024).
150. Zhou, L.; Gao, J.; Li, D.; Shum, H.Y. The Design and Implementation of XiaoIce, an Empathetic Social Chatbot. Comput. Linguist.; 2020; 46, pp. 53-93. [DOI: https://dx.doi.org/10.1162/coli_a_00368]
151. Rovira, A.; Slater, M. Reinforcement Learning as a tool to make people move to a specific location in Immersive Virtual Reality. Int. J. Hum. Comput. Stud.; 2017; 98, pp. 89-94. [DOI: https://dx.doi.org/10.1016/j.ijhcs.2016.10.007]
152. Monaro, M.; Barakova, E.; Navarin, N. Editorial Special Issue Interaction with Artificial Intelligence Systems: New Human-Centered Perspectives and Challenges. IEEE Trans. Hum. Mach. Syst.; 2022; 52, pp. 326-331. [DOI: https://dx.doi.org/10.1109/THMS.2022.3172516]
153. Otoum, Y.; Gottimukkala, N.; Kumar, N.; Nayak, A. Machine Learning in Metaverse Security: Current Solutions and Future Challenges. ACM Comput Surv.; 2024; 56, 215. [DOI: https://dx.doi.org/10.1145/3654663]
154. Xu, M.; Niyato, D.; Chen, J.; Zhang, H.; Kang, J.; Xiong, Z.; Mao, S.; Han, Z. Generative AI-Empowered Simulation for Autonomous Driving in Vehicular Mixed Reality Metaverses. IEEE J. Sel. Top. Signal Process.; 2023; 17, pp. 1064-1079. [DOI: https://dx.doi.org/10.1109/JSTSP.2023.3293650]
155. Xu, F.; Zhang, J.; Gao, C.; Feng, J.; Li, Y. Urban Generative Intelligence (UGI): A Foundational Platform for Agents in Embodied City Environment. arXiv; 2023; arXiv: 2312.11813
156. Hong, J.; Lee, Y.; Kim, D.H.; Choi, D.; Yoon, Y.J.; Lee, G.C.; Lee, Z.; Kim, J. A Context-Aware Onboarding Agent for Metaverse Powered by Large Language Models. Proceedings of the 2024 ACM Designing Interactive Systems Conference (DIS); Copenhagen, Denmark, 1–5 July 2024; pp. 1857-1874. [DOI: https://dx.doi.org/10.1145/3643834.3661579]
157. Huang, R.; Zhao, Z.; Liu, H.; Liu, J.; Cui, C.; Ren, Y. ProDiff: Progressive Fast Diffusion Model for High-Quality Text-to-Speech. Proceedings of the MM 2022—30th ACM International Conference on Multimedia; Lisboa, Portugal, 10–14 October 2022; pp. 2595-2605. [DOI: https://dx.doi.org/10.1145/3503161.3547855]
158. Salloum, S.A.; Bettayeb, A.; Salloum, A.; Aburayya, A.; Khadragy, S.; Hamoudi, R.; Alfaisal, R. Novel machine learning based approach for analysing the adoption of metaverse in medical training: A UAE case study. Inform. Med. Unlocked; 2023; 42, 101354. [DOI: https://dx.doi.org/10.1016/j.imu.2023.101354]
159. Allam, Z.; Dhunny, Z.A. On big data, artificial intelligence and smart cities. Cities; 2019; 89, pp. 80-91. [DOI: https://dx.doi.org/10.1016/j.cities.2019.01.032]
160. Chen, W.; An, J.; Li, R.; Fu, L.; Xie, G.; Bhuiyan, M.Z.; Li, K. A novel fuzzy deep-learning approach to traffic flow prediction with uncertain spatial–temporal data features. Future Gener. Comput. Syst.; 2018; 89, pp. 78-88. [DOI: https://dx.doi.org/10.1016/j.future.2018.06.021]
161. Guo, K.; Hu, Y.; Qian, Z.; Liu, H.; Zhang, K.; Sun, Y.; Gao, J.; Yin, B. Optimized Graph Convolution Recurrent Neural Network for Traffic Prediction. IEEE Trans. Intell. Transp. Syst.; 2021; 22, pp. 1138-1149. [DOI: https://dx.doi.org/10.1109/TITS.2019.2963722]
162. Wang, D.; Guo, D.; Zhang, H. Spatial temporal data visualization in emergency management: A view from data-driven decision. Proceedings of the 3rd ACM SIGSPATIAL International Workshop on the Use of GIS in Emergency Management (EM-GIS); Redondo Beach, CA, USA, 7 November 2017; [DOI: https://dx.doi.org/10.1145/3152465.3152473]
163. Chakareski, J. Aerial UAV-IoT sensing for ubiquitous immersive communication and virtual human teleportation. Proceedings of the 2017 IEEE Conference on Computer Communications Workshops; Atlanta, GA, USA, 1–4 May 2017; pp. 718-723. [DOI: https://dx.doi.org/10.1109/INFCOMW.2017.8116465]
164. Jung, T.; Dieck, M.C.T.; Lee, H.; Chung, N. Tracking Tourist Spatial-Temporal Behavior in Urban Places, A Methodological Overview and GPS Case Study. Proceedings of the Information and Communication Technologies in Tourism 2016: The International Conference; Bilbao, Spain, 2–5 February 2016; pp. 481-494. [DOI: https://dx.doi.org/10.1007/978-3-319-28231-2_35]
165. Beněs, K.; Baskar, M.K.; Burget, L. Residual memory networks in language modeling: Improving the reputation of feed-forward networks. Proceedings of the Annual Conference of the International Speech Communication Association; Stockholm, Sweden, 20–24 August 2017; Volume 2017, pp. 284-288. [DOI: https://dx.doi.org/10.21437/INTERSPEECH.2017-1442]
166. Yang, W.; Liew, Z.Q.; Lim, W.Y.; Xiong, Z.; Niyato, D.; Chi, X.; Cao, X.; Letaief, K.B. Semantic Communication Meets Edge Intelligence. IEEE Wirel. Commun.; 2022; 29, pp. 28-35. [DOI: https://dx.doi.org/10.1109/MWC.004.2200050]
167. Peng, X.; Qin, Z.; Huang, D.; Tao, X.; Lu, J.; Liu, G.; Pan, C. A Robust Deep Learning Enabled Semantic Communication System for Text. Proceedings of the IEEE Global Communications Conference; Rio de Janeiro, Brazil, 4–8 December 2022; pp. 2704-2709. [DOI: https://dx.doi.org/10.1109/GLOBECOM48099.2022.10000901]
168. Weng, Z.; Qin, Z.; Tao, X.; Pan, C.; Liu, G.; Li, G.Y. Deep Learning Enabled Semantic Communications with Speech Recognition and Synthesis. IEEE Trans. Wirel. Commun.; 2023; 22, pp. 6227-6240. [DOI: https://dx.doi.org/10.1109/TWC.2023.3240969]
169. Chowdhary, C.L.; Alazab, M.; Chaudhary, A.; Hakak, S.; Gadekallu, T.R. Computer Vision and Recognition Systems Using Machine and Deep Learning Approaches: Fundamentals, technologies and applications. Computer Vision and Recognition Systems Using Machine and Deep Learning Approaches: Fundamentals, Technologies and Applications; Institution of Engineering and Technology: Stevenage, UK, January 2021; pp. 1-483. [DOI: https://dx.doi.org/10.1049/PBPC042E]
170. Ramalingam, M.; Selvi, G.C.; Victor, N.; Chengoden, R.; Bhattacharya, S.; Maddikunta, P.K.; Lee, D.; Piran, M.J.; Khare, N.; Yenduri, G. et al. A Comprehensive Analysis of Blockchain Applications for Securing Computer Vision Systems. IEEE Access; 2023; 11, pp. 107309-107330. [DOI: https://dx.doi.org/10.1109/ACCESS.2023.3319089]
171. Fangbemi, A.S.; Lu, Y.F.; Xu, M.Y.; Luo, X.W.; Rolland, A.; Raissi, C. ZooBuilder: 2D and 3D Pose Estimation for Quadrupeds Using Synthetic Data. arXiv; 2020; arXiv: 2009.05389
172. Lo, S.C.; Tsai, H.H. Design of 3D Virtual Reality in the Metaverse for Environmental Conservation Education Based on Cognitive Theory. Sensors; 2022; 22, 8329. [DOI: https://dx.doi.org/10.3390/s22218329] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36366028]
173. Kuru, K. MetaOmniCity: Toward Immersive Urban Metaverse Cyberspaces Using Smart City Digital Twins. IEEE Access; 2023; 11, pp. 43844-43868. [DOI: https://dx.doi.org/10.1109/ACCESS.2023.3272890]
174. Qu, Q.; Hatami, M.; Xu, R.; Nagothu, D.; Chen, Y.; Li, X.; Blasch, E.; Ardiles-Cruz, E.; Chen, G. The Microverse: A Task-Oriented Edge-Scale Metaverse. Future Internet; 2024; 16, 60. [DOI: https://dx.doi.org/10.3390/fi16020060]
175. Wang, J.; Chen, Y.; Ji, X.; Dong, Z.; Gao, M.; Lai, C.S. Metaverse Meets Intelligent Transportation System: An Efficient and Instructional Visual Perception Framework. IEEE Trans. Intell. Transp. Syst.; 2024; 22, 100954. [DOI: https://dx.doi.org/10.1109/TITS.2024.3398586]
176. Boje, C.; Guerriero, A.; Kubicki, S.; Rezgui, Y. Towards a semantic Construction Digital Twin: Directions for future research. Autom. Constr.; 2020; 114, 103179. [DOI: https://dx.doi.org/10.1016/j.autcon.2020.103179]
177. Zhao, L.; Zhang, H.; Wang, Q.; Wang, H. Digital-Twin-Based Evaluation of Nearly Zero-Energy Building for Existing Buildings Based on Scan-to-BIM. Advances in Civil. Eng.; 2021; 2021, 6638897. [DOI: https://dx.doi.org/10.1155/2021/6638897]
178. Austin, M.; Delgoshaei, P.; Coelho, M.; Heidarinejad, M. Architecting Smart City Digital Twins: Combined Semantic Model and Machine Learning Approach. J. Manag. Eng.; 2020; 36, 04020026. [DOI: https://dx.doi.org/10.1061/(ASCE)ME.1943-5479.0000774]
179. Sahal, R.; Alsamhi, S.H.; Brown, K.N.; O’shea, D.; McCarthy, C.; Guizani, M. Blockchain-Empowered Digital Twins Collaboration: Smart Transportation Use Case. Machines; 2021; 9, 193. [DOI: https://dx.doi.org/10.3390/machines9090193]
180. Hu, X.; Zhou, Y.; Shi, Q. Method for 2D-3D Registration under Inverse Depth and Structural Semantic Constraints for Digital Twin City. Appl. Sci.; 2022; 12, 8543. [DOI: https://dx.doi.org/10.3390/app12178543]
181. Li, X.; Wang, L.; Zhu, C.; Liu, Z. Framework for manufacturing-tasks semantic modelling and manufacturing-resource recommendation for digital twin shop-floor. J. Manuf. Syst.; 2021; 58, pp. 281-292. [DOI: https://dx.doi.org/10.1016/j.jmsy.2020.08.003]
182. Yang, Q.; Zhao, Y.; Huang, H.; Xiong, Z.; Kang, J.; Zheng, Z. Fusing Blockchain and AI With Metaverse: A Survey. IEEE Open J. Comput. Soc.; 2022; 3, pp. 122-136. [DOI: https://dx.doi.org/10.1109/OJCS.2022.3188249]
183. Lambert, N. Beyond NFTs: A Possible Future for Digital Art. ITNOW; 2021; 63, pp. 8-10. [DOI: https://dx.doi.org/10.1093/itnow/bwab066]
184. Blockchain: Rebalancing & Amplifying the Power of AI and Machine Learning (ML) | by Jorden Woods | JustStable | Medium. Available online: https://medium.com/crypto-oracle/blockchain-rebalancing-amplifying-the-power-of-ai-and-machine-learning-ml-af95616e9ad9 (accessed on 15 August 2024).
185. Liu, Y.; Zhang, Y.; Su, S.; Zhang, L.; Du, X.; Guizani, M.; Tian, Z. BlockSC: A Blockchain Empowered Spatial Crowdsourcing Service in Metaverse while Preserving User Location Privacy. IEEE J. Sel. Areas Commun.; 2024; 42, pp. 880-892. [DOI: https://dx.doi.org/10.1109/JSAC.2023.3345416]
186. Khan, L.U.; Yaqoob, I.; Tran, N.H.; Kazmi, S.M.A.; Dang, T.N.; Hong, C.S. Edge-Computing-Enabled Smart Cities: A Comprehensive Survey. IEEE Internet Things J.; 2020; 7, pp. 10200-10232. [DOI: https://dx.doi.org/10.1109/JIOT.2020.2987070]
187. Lv, Z.; Qiao, L.; Li, Y.; Yuan, Y.; Wang, F.Y. BlockNet: Beyond reliable spatial Digital Twins to Parallel Metaverse. Patterns; 2022; 3, 100468. [DOI: https://dx.doi.org/10.1016/j.patter.2022.100468]
188. Cliff, D.; Rollins, M. Methods Matter: A Trading Agent with No Intelligence Routinely Outperforms AI-Based Traders. November 2020. Available online: https://arxiv.org/abs/2011.14346v1 (accessed on 15 August 2024).
189. Lu, E.; Miller, J.; Pereira, N.; Rowe, A. FlasH: Video-embeddable AR anchors for live events. Proceedings of the—2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR); Bari, Italy, 4–8 October 2021; pp. 489-497. [DOI: https://dx.doi.org/10.1109/ISMAR52148.2021.00066]
190. Xie, H.; Qin, Z. A Lite Distributed Semantic Communication System for Internet of Things. IEEE J. Sel. Areas Commun.; 2021; 39, pp. 142-153. [DOI: https://dx.doi.org/10.1109/JSAC.2020.3036968]
191. Shi, G.; Xiao, Y.; Li, Y.; Xie, X. From Semantic Communication to Semantic-Aware Networking: Model, Architecture, and Open Problems. IEEE Commun. Mag.; 2021; 59, pp. 44-50. [DOI: https://dx.doi.org/10.1109/MCOM.001.2001239]
192. Rubí, J.N.S.; de Lira Gondim, P.R. IoT-based platform for environment data sharing in Smart Cities. Int. J. Commun. Syst.; 2023; 34, pp. 64-71. [DOI: https://dx.doi.org/10.1109/MWC.018.2100716]
193. Solmaz, G.; Wu, F.J.; Cirillo, F.; Kovacs, E.; Santana, J.R.; Sánchez, L.; Sotres, P.; Munoz, L. Toward Understanding Crowd Mobility in Smart Cities through the Internet of Things. IEEE Commun. Mag.; 2019; 57, pp. 40-46. [DOI: https://dx.doi.org/10.1109/MCOM.2019.1800611]
194. Dev, K.; Khowaja, S.A.; Sharma, P.K.; Chowdhry, B.S.; Tanwar, S.; Fortino, G. DDI: A Novel Architecture for Joint Active User Detection and IoT Device Identification in Grant-Free NOMA Systems for 6G and beyond Networks. IEEE Internet Things J.; 2022; 9, pp. 2906-2917. [DOI: https://dx.doi.org/10.1109/JIOT.2021.3095255]
195. She, C.; Dong, R.; Gu, Z.; Hou, Z.; Li, Y.; Hardjawana, W.; Yang, C.; Song, L.; Vucetic, B. Deep Learning for Ultra-Reliable and Low-Latency Communications in 6G Networks. IEEE Netw.; 2020; 34, pp. 219-225. [DOI: https://dx.doi.org/10.1109/MNET.011.1900630]
196. Alsenwi, M.; Tran, N.H.; Bennis, M.; Pandey, S.R.; Bairagi, A.K.; Hong, C.S. Intelligent Resource Slicing for eMBB and URLLC Coexistence in 5G and Beyond: A Deep Reinforcement Learning Based Approach. IEEE Trans. Wirel. Commun.; 2021; 20, pp. 4585-4600. [DOI: https://dx.doi.org/10.1109/TWC.2021.3060514]
197. Gu, B.; Zhang, X.; Lin, Z.; Alazab, M. Deep Multiagent Reinforcement-Learning-Based Resource Allocation for Internet of Controllable Things. IEEE Internet Things J.; 2021; 8, pp. 3066-3074. [DOI: https://dx.doi.org/10.1109/JIOT.2020.3023111]
198. Cheng, N.; Lyu, F.; Quan, W.; Zhou, C.; He, H.; Shi, W.; Shen, X. Space/Aerial-Assisted Computing Offloading for IoT Applications: A Learning-Based Approach. IEEE J. Sel. Areas Commun.; 2019; 37, pp. 1117-1129. [DOI: https://dx.doi.org/10.1109/JSAC.2019.2906789]
199. Wang, H.; Wang, Z.; Chen, D.; Liu, Q.; Ke, H.; Han, K.K.T. Metamobility: Connecting Future Mobility with the Metaverse. IEEE Veh. Technol. Mag.; 2023; 18, pp. 69-79. [DOI: https://dx.doi.org/10.1109/MVT.2023.3263330]
200. Aslam, A.M.; Chaudhary, R.; Bhardwaj, A.; Budhiraja, I.; Kumar, N.; Zeadally, S. Metaverse for 6G and Beyond: The Next Revolution and Deployment Challenges. IEEE Internet Things Mag.; 2023; 6, pp. 32-39. [DOI: https://dx.doi.org/10.1109/IOTM.001.2200248]
201. Giordani, M.; Polese, M.; Mezzavilla, M.; Rangan, S.; Zorzi, M. Toward 6G Networks: Use Cases and Technologies. IEEE Commun. Mag.; 2020; 58, pp. 55-61. [DOI: https://dx.doi.org/10.1109/MCOM.001.1900411]
202. Osorio, D.P.; Ahmad, I.; Sánchez, J.D.; Gurtov, A.; Scholliers, J.; Kutila, M.; Porambage, P. Towards 6G-Enabled Internet of Vehicles: Security and Privacy. IEEE Open J. Commun. Soc.; 2022; 3, pp. 82-105. [DOI: https://dx.doi.org/10.1109/OJCOMS.2022.3143098]
203. Tunze, G.B.; Huynh-The, T.; Lee, J.M.; Kim, D.S. Sparsely Connected CNN for Efficient Automatic Modulation Recognition. IEEE Trans. Veh. Technol.; 2020; 69, pp. 15557-15568. [DOI: https://dx.doi.org/10.1109/TVT.2020.3042638]
204. Ullo, S.L.; Sinha, G.R. Advances in Smart Environment Monitoring Systems Using IoT and Sensors. Sensors; 2020; 20, 3113. [DOI: https://dx.doi.org/10.3390/s20113113]
205. Saad, M.; Ahmad, M.B.; Asif, M.; Khan, M.K.; Mahmood, T.; Mahmood, M.T. Blockchain-Enabled VANET for Smart Solid Waste Management. IEEE Access; 2023; 11, pp. 5679-5700. [DOI: https://dx.doi.org/10.1109/ACCESS.2023.3235017]
206. Yu, Y.; Chen, Y.; Wei, R.; Zhang, X.; Li, K.; Wan, D. A Deep Learning-Based Multi-model Ensemble Method for Hydrological Forecasting. Proceedings of the 2022 IEEE SmartWorld, Ubiquitous Intelligence and Computing, Autonomous and Trusted Vehicles, Scalable Computing and Communications, Digital Twin, Privacy Computing, Metaverse, SmartWorld/UIC/ATC/ScalCom/DigitalTwin/PriComp/Metaverse 2022; Haikou, China, 15–18 December 2022; pp. 1245-1251. [DOI: https://dx.doi.org/10.1109/SMARTWORLD-UIC-ATC-SCALCOM-DIGITALTWIN-PRICOMP-METAVERSE56740.2022.00184]
207. Gu, J.; Wang, J.; Guo, X.; Liu, G.; Qin, S.; Bi, Z. A Metaverse-Based Teaching Building Evacuation Training System with Deep Reinforcement Learning. IEEE Trans. Syst. Man. Cybern. Syst.; 2023; 53, pp. 2209-2219. [DOI: https://dx.doi.org/10.1109/TSMC.2022.3231299]
208. Mao, W.; Wang, W.; Jiao, L.; Zhao, S.; Liu, A. Modeling air quality prediction using a deep learning approach: Method optimization and evaluation. Sustain. Cities Soc.; 2021; 65, 102567. [DOI: https://dx.doi.org/10.1016/j.scs.2020.102567]
209. Nunez-Cacho, P.; Maqueira-Marin, J.M.; Rata, B.M.; Molina, V.-M. Building a model for the predictive improvement of air quality in Circular Smart Cities. Proceedings of the ISC2 2022—8th IEEE International Smart Cities Conference; Pafos, Cyprus, 26–29 September 2022; [DOI: https://dx.doi.org/10.1109/ISC255366.2022.9922373]
210. Haq, M.A.; Author, C. SMOTEDNN: A Novel Model for Air Pollution Forecasting and AQI Classification. Comput. Mater. Contin.; 2022; 71, pp. 1403-1425. [DOI: https://dx.doi.org/10.32604/cmc.2022.021968]
211. Mao, W.L.; Chen, W.C.; Wang, C.T.; Lin, Y.H. Recycling waste classification using optimized convolutional neural network. Resour. Conserv. Recycl.; 2021; 164, 105132. [DOI: https://dx.doi.org/10.1016/j.resconrec.2020.105132]
212. Han, H.; Fan, X.; Li, F. Prototype Enhancement-Based Incremental Evolution Learning for Urban Garbage Classification. IEEE Trans. Artif. Intell.; 2024; 5, pp. 398-411. [DOI: https://dx.doi.org/10.1109/TAI.2023.3250207]
213. Jiang, L.; Shi, J.; Wang, C.; Pan, Z. Intelligent control of building fire protection system using digital twins and semantic web technologies. Autom. Constr.; 2023; 147, 104728. [DOI: https://dx.doi.org/10.1016/j.autcon.2022.104728]
214. Eneyew, D.D.; Capretz, M.A.M.; Bitsuamlak, G.T. Toward Smart-Building Digital Twins: BIM and IoT Data Integration. IEEE Access; 2022; 10, pp. 130487-130506. [DOI: https://dx.doi.org/10.1109/ACCESS.2022.3229370]
215. Das, M.; Pratama, M.; Ghosh, S.K. SARDINE: A Self-Adaptive Recurrent Deep Incremental Network Model for Spatio-Temporal Prediction of Remote Sensing Data. ACM Trans. Spat. Algorithms Syst.; 2020; 6, [DOI: https://dx.doi.org/10.1145/3380972]
216. Du, S.; Li, T.; Yang, Y.; Horng, S.J. Deep Air Quality Forecasting Using Hybrid Deep Learning Framework. IEEE Trans. Knowl. Data Eng.; 2021; 33, pp. 2412-2424. [DOI: https://dx.doi.org/10.1109/TKDE.2019.2954510]
217. Kim, E.E.; Rossi, R.M.; Lorenzo, B.C.; Magoutas, B.; Martinez, A.F.; Crettaz, C.; Bothos, E.; Montenegro, J.M. Water economy with smart water system in the City of Carouge. Proceedings of the 2022 IEEE International Conference on Omni-Layer Intelligent Systems (COINS); Barcelona, Spain, 1–3 August 2022; [DOI: https://dx.doi.org/10.1109/COINS54846.2022.9854990]
218. Lin, J.Y.; Lin, Y.B.; Chen, W.L.; Ng, F.L.; Yeh, J.H.; Lin, Y.W. IoT-Based Bacillus Number Prediction in Smart Turmeric Farms Using Small Data Sets. IEEE Internet Things J.; 2023; 10, pp. 5146-5157. [DOI: https://dx.doi.org/10.1109/JIOT.2022.3222283]
219. Islam, M.D.; Di, L.; Qamer, F.M.; Shrestha, S.; Guo, L.; Lin, L.; Mayer, T.J.; Phalke, A.R. Rapid Rice Yield Estimation Using Integrated Remote Sensing and Meteorological Data and Machine Learning. Remote Sens.; 2023; 15, 2374. [DOI: https://dx.doi.org/10.3390/rs15092374]
220. Wang, X.; Kang, M.; Sun, H.; De Reffye, P.; Wang, F.Y. DeCASA in AgriVerse: Parallel Agriculture for Smart Villages in Metaverses. IEEE/CAA J. Autom. Sin.; 2022; 9, pp. 2055-2062. [DOI: https://dx.doi.org/10.1109/JAS.2022.106103]
221. Decentralised Material Recovery for a Sustainable Future. Available online: https://ercim-news.ercim.eu/en138/special/decentralised-material-recovery-for-a-sustainable-future (accessed on 26 July 2024).
222. Han, J.; Liu, H.; Liu, S.; Chen, X.; Tan, N.; Chai, H.; Xiong, H. IETA: A Robust and Scalable Incremental Learning Framework for Time-of-Arrival Estimation. Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; Long Beach, CA, USA, 6–10 August 2023; Volume 23, pp. 4100-4111. [DOI: https://dx.doi.org/10.1145/3580305.3599842]
223. Chen, X.; Wang, J.; Xie, K. TrafficStream: A Streaming Traffic Flow Forecasting Framework Based on Graph Neural Networks and Continual Learning. Proceedings of the IJCAI International Joint Conference on Artificial Intelligence; Montreal, QC, Canada, 19–27 August 2021; pp. 3620-3626. [DOI: https://dx.doi.org/10.24963/ijcai.2021/498]
224. Guo, S.; Lin, Y.; Li, S.; Chen, Z.; Wan, H. Deep Spatialoral 3D Convolutional Neural Networks for Traffic Data Forecasting. IEEE Trans. Intell. Transp. Syst.; 2019; 20, pp. 3913-3926. [DOI: https://dx.doi.org/10.1109/TITS.2019.2906365]
225. Rui, Y.; Gong, Y.; Zhao, Y.; Luo, K.; Lu, W. Predicting Traffic Flow Parameters for Sustainable Highway Management: An Attention-Based EMD–BiLSTM Approach. Sustainability; 2023; 16, 190. [DOI: https://dx.doi.org/10.3390/su16010190]
226. Maipradit, A.; Moriyama, Y.; Okuro, T.; Yoshida, M.; Tachimori, N.; Akiyama, S.; Suwa, H.; Yasumoto, K. PAVEMENT: Passing Vehicle Detection System with Autonomous Incremental Learning using Camera and Vibration Data. Proceedings of the IEEE Vehicular Technology Conference; Helsinki, Finland, 19–22 June 2022; Volume 2022, [DOI: https://dx.doi.org/10.1109/VTC2022-FALL57202.2022.10012958]
227. Nan, Y.; Jiang, S.; Li, M. Large-scale Video Analytics with Cloud–Edge Collaborative Continuous Learning. ACM Trans. Sens. Netw.; 2023; 20, 14. [DOI: https://dx.doi.org/10.1145/3624478]
228. Bao, P.; Chen, Z.; Wang, J.; Dai, D.; Zhao, H. Lifelong Vehicle Trajectory Prediction Framework Based on Generative Replay. IEEE Trans. Intell. Transp. Syst.; 2023; 24, pp. 13729-13741. [DOI: https://dx.doi.org/10.1109/TITS.2023.3300545]
229. Habibi, G.; Jaipuria, N.; How, J.P. SILA: An Incremental Learning Approach for Pedestrian Trajectory Prediction. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops; Seattle, WA, USA, 14–19 June 2020.
230. Bandaragoda, T.; De Silva, D.; Kleyko, D.; Osipov, E.; Wiklund, U.; Alahakoon, D. Trajectory clustering of road traffic in urban environments using incremental machine learning in combination with hyperdimensional computing. Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference (ITSC); Auckland, New Zealand, 27–30 October 2019; pp. 1664-1670. [DOI: https://dx.doi.org/10.1109/ITSC.2019.8917320]
231. D3—An AI-based Solution to Road Defect Detection. Available online: https://ercim-news.ercim.eu/en138/r-i/d3---an-ai-based-solution-to-road-defect-detection (accessed on 28 August 2024).
232. Nuvvula, R.S.; Kumar, P.P.; Siddiqui, A.; Thamizharasan, S.; Tan, C.C.; Alubady, R.; Khan, B. Optimizing Electric Vehicle Fleet Operations with Predictive Analytics: A Renewable Energy-Centric Approach. In Proceeding of the 2024 12th International Conference on Smart Grid (icSmartGrid); Setubal, Portugal, 27–29 May 2024; pp. 309-311. [DOI: https://dx.doi.org/10.1109/ICSMARTGRID61824.2024.10578265]
233. Bellini, P.; Bilotta, S.; Collini, E.; Fanfani, M.; Nesi, P. Mobility and Transport Data for City Digital Twin Modeling and Exploitation. Proceedings of the 2023 IEEE International Smart Cities Conference (ISC2); Bucharest, Romania, 24–27 September 2023; [DOI: https://dx.doi.org/10.1109/ISC257844.2023.10293300]
234. Njoku, J.N.; Nwakanma, C.I.; Kim, D.-S. Evaluation of Spectrograms for Keyword Spotting in Control of Autonomous Vehicles for The Metaverse. Available online: https://www.researchgate.net/publication/361558505 (accessed on 12 September 2024).
235. Li, B.; Cao, D.; Tang, S.; Zhang, T.; Dong, H.; Wang, Y.; Wang, F.Y. Sharing Traffic Priorities via Cyber-Physical-Social Intelligence: A Lane-Free Autonomous Intersection Management Method in Metaverse. IEEE Trans. Syst. Man. Cybern. Syst.; 2023; 53, pp. 2025-2036. [DOI: https://dx.doi.org/10.1109/TSMC.2022.3225250]
236. Tan, L.; Yu, K.; Lin, L.; Cheng, X.; Srivastava, G.; Lin, J.C.; Wei, W. Speech Emotion Recognition Enhanced Traffic Efficiency Solution for Autonomous Vehicles in a 5G-Enabled Space-Air-Ground Integrated Intelligent Transportation System. IEEE Trans. Intell. Transp. Syst.; 2022; 23, pp. 2830-2842. [DOI: https://dx.doi.org/10.1109/TITS.2021.3119921]
237. Fan, L.; Cao, D.; Zeng, C.; Li, B.; Li, Y.; Wang, F.Y. Cognitive-Based Crack Detection for Road Maintenance: An Integrated System in Cyber-Physical-Social Systems. IEEE Trans. Syst. Man. Cybern. Syst.; 2023; 53, pp. 3485-3500. [DOI: https://dx.doi.org/10.1109/TSMC.2022.3227209]
238. Hu, T.; Wu, W.; Guo, Q.; Sun, H.; Shi, L.; Shen, X. Very short-term spatial and temporal wind power forecasting: A deep learning approach. CSEE J. Power Energy Syst.; 2020; 6, pp. 434-443. [DOI: https://dx.doi.org/10.17775/CSEEJPES.2018.00010]
239. Hong, S.H.; Lee, H.S. Robust Energy Management System with Safe Reinforcement Learning Using Short-Horizon Forecasts. IEEE Trans. Smart Grid; 2023; 14, pp. 2485-2488. [DOI: https://dx.doi.org/10.1109/TSG.2023.3240588]
240. Donatos, S.; Panagiotis, T.; Thanasis, K. Enhancing Residential Energy Efficiency and Management Through IoT and Smart Grid Integration. Available online: https://ercim-news.ercim.eu/en138/special/enhancing-residential-energy-efficiency-and-management-through-iot-and-smart-grid-integration (accessed on 27 August 2024).
241. Abid, Y.M.; Jalil, M.A.; Abd, S.K.; Sabah, H.A.; Bakri, B.I.; Maktoof, M.A.; Mohammed, A.; Adile, M.; Ahmed, M.S.; Mohammed, M.A. et al. Analysis of the Energy Efficiency of Smart City Buildings Based on Deep Learning Algorithms Using AI. Proceedings of the 2024 International Conference on Smart Systems for Electrical, Electronics, Communication and Computer Engineering (ICSSEECC); Coimbatore, India, 28–29 June 2024; pp. 672-677. [DOI: https://dx.doi.org/10.1109/ICSSEECC61126.2024.10649512]
242. Khan, N.; Khan, S.U.; Ullah, F.U.M.; Lee, M.Y.; Baik, S.W. AI-Assisted Hybrid Approach for Energy Management in IoT-Based Smart Microgrid. IEEE Internet Things J.; 2023; 10, pp. 18861-18875. [DOI: https://dx.doi.org/10.1109/JIOT.2023.3293800]
243. He, X.; Ai, Q.; Wang, J.; Tao, F.; Pan, B.; Qiu, R.; Yang, B. Situation Awareness of Energy Internet of Things in Smart City Based on Digital Twin: From Digitization to Informatization. IEEE Internet Things J.; 2023; 10, pp. 7439-7458. [DOI: https://dx.doi.org/10.1109/JIOT.2022.3203823]
244. Qayyum, F.; Iqbal, N.; Jamil, H.; Rizwan, A.; Khan, A.N.; Ahmad, R.; Khan, S.; Kim, D.H. Energy Trading Framework Based on IoT and Digital Twin for Nanogrid Environment. Proceedings of the—2023 IEEE International Conference on Metaverse Computing, Networking and Applications (MetaCom); Kyoto, Japan, 26–28 June 2023; pp. 749-755. [DOI: https://dx.doi.org/10.1109/METACOM57706.2023.00142]
245. Ma, W.; Liu, M.; Hong, G.; Yang, S.; Deng, R. Grid-Metaverse: The Path from Digital Twins and Prototype Tests on DC Microgrids. Proceedings of the—2023 IEEE International Conference on Metaverse Computing, Networking and Applications (MetaCom); Kyoto, Japan, 26–28 June 2023; pp. 290-296. [DOI: https://dx.doi.org/10.1109/METACOM57706.2023.00059]
246. Farooq, J.; Bazaz, M.A. A novel adaptive deep learning model of Covid-19 with focus on mortality reduction strategies. Chaos Solitons Fractals; 2020; 138, 110148. [DOI: https://dx.doi.org/10.1016/j.chaos.2020.110148]
247. Camargo, E.; Aguilar, J.; Quintero, Y.; Rivas, F.; Ardila, D. An incremental learning approach to prediction models of SEIRD variables in the context of the COVID-19 pandemic. Health Technol.; 2022; 12, pp. 867-877. [DOI: https://dx.doi.org/10.1007/s12553-022-00668-5]
248. Wang, M.; Yu, D.; He, W.; Yue, P.; Liang, Z. Domain-incremental learning for fire detection in space-air-ground integrated observation network. Int. J. Appl. Earth Obs. Geoinf.; 2023; 118, 103279. [DOI: https://dx.doi.org/10.1016/j.jag.2023.103279]
249. Kim, J.; Hong, S.; Song, M.; Kim, K. Visual Attention and Pulmonary VR Training System for Children with Attention Deficit Hyperactivity Disorder. IEEE Access; 2024; 12, pp. 53739-53751. [DOI: https://dx.doi.org/10.1109/ACCESS.2024.3387065]
250. Zanitti, M.; Ferens, M.; Ferrarin, A.; Trovò, F.; Miskovic, V.; Prelaj, A.; Shen, M.; Kosta, S. MetaLung: Towards a Secure Architecture for Lung Cancer Patient Care on the Metaverse. Proceedings of the—2023 IEEE International Conference on Metaverse Computing, Networking and Applications (MetaCom); Kyoto, Japan, 26–28 June 2023; pp. 201-208. [DOI: https://dx.doi.org/10.1109/METACOM57706.2023.00047]
251. Qiao, S.; Pang, S.; Sun, Y.; Luo, G.; Yin, W.; Zhao, Y.; Pan, S.; Lv, Z. SPReCHD: Four-Chamber Semantic Parsing Network for Recognizing Fetal Congenital Heart Disease in Medical Metaverse. IEEE J. Biomed. Health Inform.; 2024; 28, pp. 3672-3682. [DOI: https://dx.doi.org/10.1109/JBHI.2022.3218577]
252. Murala, D.K.; Panda, S.K.; Dash, S.P. MedMetaverse: Medical Care of Chronic Disease Patients and Managing Data Using Artificial Intelligence, Blockchain, and Wearable Devices State-of-the-Art Methodology. IEEE Access; 2023; 11, pp. 138954-138985. [DOI: https://dx.doi.org/10.1109/ACCESS.2023.3340791]
253. Tian, W.; Lu, Y.; Yu, J.; Fan, J.; Tang, P.; Zhang, L. A Privacy-Preserving Framework for Mental Health Chatbots Based on Confidential Computing. Proceedings of the—2022 IEEE SmartWorld, Ubiquitous Intelligence and Computing, Autonomous and Trusted Vehicles, Scalable Computing and Communications, Digital Twin, Privacy Computing, Metaverse, SmartWorld/UIC/ATC/ScalCom/DigitalTwin/PriComp/Metaverse 2022; Haikou, China, 15–18 December 2022; pp. 1119-1124. [DOI: https://dx.doi.org/10.1109/SMARTWORLD-UIC-ATC-SCALCOM-DIGITALTWIN-PRICOMP-METAVERSE56740.2022.00160]
254. Zikas, P.; Protopsaltis, A.; Lydatakis, N.; Kentros, M.; Geronikolakis, S.; Kateros, S.; Kamarianakis, M.; Evangelou, G.; Filippidis, A.; Grigoriou, E. et al. MAGES 4.0: Accelerating the World’s Transition to VR Training and Democratizing the Authoring of the Medical Metaverse. IEEE Comput. Graph. Appl.; 2023; 43, pp. 43-56. [DOI: https://dx.doi.org/10.1109/MCG.2023.3242686]
255. Zikas, P.; Kamarianakis, M.; Kartsonaki, I.; Lydatakis, N.; Kateros, S.; Kentros, M.; Geronikolakis, E.; Evangelou, G.; Apostolou, A.; Catilo, P.A. et al. COVID-19—VR Strikes Back: Innovative medical VR training. ACM SIGGRAPH 2021 Immersive Pavilion, SIGGRAPH 2021; Association for Computing Machinery: New York, NY, USA, 2021; [DOI: https://dx.doi.org/10.1145/3450615.3464546]
256. Singh, R.; Nandi, S.; Nandi, S.K. An Implementation and Analysis of Zero Knowledge Based E-Voting Solution with Proof of Vote on Public Ethereum Blockchain. Proceedings of the—2023 IEEE International Conference on Metaverse Computing, Networking and Applications (MetaCom); Kyoto, Japan, 26–28 June 2023; pp. 423-427. [DOI: https://dx.doi.org/10.1109/METACOM57706.2023.00081]
257. Decision Support System for Inclusive Smart City Design: Baby Boomers as a Use Case. Available online: https://ercim-news.ercim.eu/en138/special/decision-support-system-for-inclusive-smart-city-design-baby-boomers-as-a-use-case (accessed on 26 July 2024).
258. Ding, W.; Li, J.; Qin, R.; Kozma, R.; Wang, F.Y. A New Architecture and Mechanism for Decentralized Science MetaMarkets. IEEE Trans. Syst. Man. Cybern. Syst.; 2023; 53, pp. 5321-5330. [DOI: https://dx.doi.org/10.1109/TSMC.2023.3266223]
259. Virtual Singapore—Building a 3D-Empowered Smart Nation—Geospatial World. GW Prime. Available online: https://geospatialworld.net/prime/case-study/national-mapping/virtual-singapore-building-a-3d-empowered-smart-nation/ (accessed on 21 November 2024).
260. Official Release of Metaverse Seoul—Seoul Metropolitan Government. Available online: https://english.seoul.go.kr/official-release-of-metaverse-seoul/ (accessed on 21 November 2024).
261. Abraham, M.; Aithal, H.; Mohan, K. Blockchain and Collaborative Intelligence based next generation Smart Toll Application. Proceedings of the 2020 2nd Conference on Blockchain Research and Applications for Innovative Networks and Services, BRAINS 2020; Paris, France, 28–30 September 2020; pp. 206-207. [DOI: https://dx.doi.org/10.1109/BRAINS49436.2020.9223296]
262. Zhang, T.; Ding, J.; Zeng, C.; Guan, K.; Liu, Y.; Zhao, C.; Chai, T. IHPPPVis: Interactive Visual Analytics Approach for Production Performance Monitoring of Heavy-Plate Production Process. IEEE Trans. Cybern.; 2024; 54, pp. 3864-3877. [DOI: https://dx.doi.org/10.1109/TCYB.2024.3387129]
263. Hou, J.; Chen, G.; Li, Z.; He, W.; Gu, S.; Knoll, A.; Jiang, C. Hybrid Residual Multiexpert Reinforcement Learning for Spatial Scheduling of High-Density Parking Lots. IEEE Trans. Cybern.; 2024; 54, pp. 2771-2783. [DOI: https://dx.doi.org/10.1109/TCYB.2023.3312647]
264. Sun, F.; Liu, N.; Wang, X.; Sun, R.; Miao, S.; Kang, Z.; Fang, B.; Liu, H.; Zhao, Y.; Huang, H. Digital-Twin-Assisted Skill Learning for 3C Assembly Tasks. IEEE Trans. Cybern.; 2024; 54, pp. 3852-3863. [DOI: https://dx.doi.org/10.1109/TCYB.2024.3368148]
265. Kim, J.H.; Kim, M.; Park, M.; Yoo, J. Immersive interactive technologies and virtual shopping experiences: Differences in consumer perceptions between augmented reality (AR) and virtual reality (VR). Telemat. Inform.; 2023; 77, 101936. [DOI: https://dx.doi.org/10.1016/j.tele.2022.101936]
266. Wang, E.; Attard, S.; Everingham, Y.; Philippa, B.; Xiang, W. Smarter Irrigation Management in the Sugarcane Farming System Using Internet of Things. 2018; Available online: https://www.assct.com.au/component/assct/search-result?search_cat=title&filter_search=Smarter%20irrigation%20management%20in%20the%20sugarcane%20far&publisher=any&Itemid=0 (accessed on 30 August 2024).
267. Njoku, J.N.; Nwakanma, C.I.; Amaizu, G.C.; Kim, D.S. Prospects and challenges of Metaverse application in data-driven intelligent transportation systems. IET Intell. Transp. Syst.; 2023; 17, pp. 1-21. [DOI: https://dx.doi.org/10.1049/itr2.12252]
268. Batista, E.; Lopez-Aguilar, P.; Solanas, A. Smart Health in the 6G Era: Bringing Security to Future Smart Health Services. IEEE Commun. Mag.; 2024; 62, pp. 74-80. [DOI: https://dx.doi.org/10.1109/MCOM.019.2300122]
269. Mahmoud, M.; Perez-Ortiz, M.; Bokhari, S.A.A.; Myeong, S. Use of Artificial Intelligence in Smart Cities for Smart Decision-Making: A Social Innovation Perspective. Sustainability; 2022; 14, 620. [DOI: https://dx.doi.org/10.3390/su14020620]
270. Khansari, N.; Silverman, B.G.; Du, Q.; Waldt, J.B.; Braham, W.W.; Lee, J.M. An agent-based decision tool to explore urban climate & smart city possibilities. Proceedings of the 11th Annual IEEE International Systems Conference, SysCon 2017; Montreal, QC, Canada, 24–27 April 2017; [DOI: https://dx.doi.org/10.1109/SYSCON.2017.7934788]
271. Deep Learning Computer Vision Algorithms, Customer Engagement Tools, and Virtual Marketplace Dynamics Data in the Metaverse Economy. J. Self-Gov. Manag. Econ.; 2022; 10, 37. [DOI: https://dx.doi.org/10.22381/jsme10220223]
272. Mohamed, G.; Zahra, F.F.; Najwa, Z.; Soumaya, O.; Yassine, K.; Chakir, A.; Mohamed, A. Enhancing Immersive Virtual Shopping Experiences in the Retail Metaverse Through Visual Analytics, Cognitive Artificial Intelligence Techniques, Blockchain-Based Digital Assets, and Immersive Simulations: A Systematic Literature Review. Eng. Appl. Artif. Intell; 2024; pp. 305-318. [DOI: https://dx.doi.org/10.1007/978-3-031-50300-9_17]
273. Chatzopoulou, I.; Tsoutsa, P.; Fitsilis, P. How Metaverse is Affecting Smart Cities Economy. Proceedings of the 27th Pan-Hellenic Conference on Progress in Computing and Informatics; Lamia, Greece, 24–26 November 2023; pp. 254-259. [DOI: https://dx.doi.org/10.1145/3635059.3635099]
274. Zhaki, A. More Than 100 Solutions Developed as Part of Generative AI Initiative | The Straits Times. Available online: https://www.straitstimes.com/tech/tech-news/more-than-100-solutions-developed-as-part-of-generative-ai-initiative?close=true (accessed on 21 November 2024).
275. Aydta, H.; Acerob, J.A.; Ivanchevc, J.; Nevatc, I.; Bennya, J.; Adeliaa, A.S.; Niffelera, M.; Wonga, M.L.; Zozayaa, A.; Orehounig, K. A Digital Urban Climate Twin for Singapore. 2024; Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5002636 (accessed on 21 November 2024).
276. The First Robotaxis are Being Tested in China’s Major Cities—World Economic Forum. Available online: https://www.weforum.org/stories/2021/10/robotaxis-expected-driverless-china-tier-one-cities/ (accessed on 21 November 2024).
277. Digital Twin to Help State of New Mexico Go Carbon Neutral—SustMeme. Available online: https://sustmeme.com/2020/10/13/news-digital-twin-to-help-state-of-new-mexico-go-carbon-neutral/ (accessed on 21 November 2024).
278. Lisbon’s City Scale Digital Twins for Flood Resilience—Geospatial World. Available online: https://geospatialworld.net/prime/case-study/aec/lisbons-city-scale-digital-twins-for-flood-resilience-2/ (accessed on 21 November 2024).
279. Xu, H.; Omitaomu, F.; Sabri, S.; Zlatanova, S.; Li, X.; Song, Y. Leveraging Generative AI for Urban Digital Twins: A Scoping Review on the Autonomous Generation of Urban Data, Scenarios, Designs, and 3D City Models for Smart City Advancement. May 2024. Available online: https://arxiv.org/abs/2405.19464v2 (accessed on 19 November 2024).
280. Sai, S.; Mittal, U.; Chamola, V. DMDAT: Diffusion Model-based Data Augmentation Technique for Vision-based Accident Detection in Vehicular Networks. IEEE Trans. Veh. Technol.; 2024; [DOI: https://dx.doi.org/10.1109/TVT.2024.3491170]
281. Dabboussi, A.H.; Jammal, M. Traffic Data Augmentation Using GANs for ITS. Proceedings of the 2024 20th International Conference on Distributed Computing in Smart Systems and the Internet of Things, DCOSS-IoT 2024; Abu Dhabi, United Arab Emirates, 29 April–1 May 2024; pp. 66-73. [DOI: https://dx.doi.org/10.1109/DCOSS-IOT61029.2024.00020]
282. Javed, A.R.; Ahmed, W.; Pandya, S.; Maddikunta, P.K.R.; Alazab, M.; Gadekallu, T.R. A Survey of Explainable Artificial Intelligence for Smart Cities. Electronics; 2023; 12, 1020. [DOI: https://dx.doi.org/10.3390/electronics12041020]
283. Yigitcanlar, T.; Corchado, J.M.; Mehmood, R.; Li, R.Y.M.; Mossberger, K.; Desouza, K. Responsible Urban Innovation with Local Government Artificial Intelligence (AI): A Conceptual Framework and Research Agenda. J. Open Innov. Technol. Mark. Complex.; 2021; 7, 71. [DOI: https://dx.doi.org/10.3390/joitmc7010071]
284. Liu, Z.; Meyendorf, N.; Mrad, N. The role of data fusion in predictive maintenance using digital twin. AIP Conf. Proc.; 2018; 1949, 20023. [DOI: https://dx.doi.org/10.1063/1.5031520]
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Ding, Jianguo 2
; Huansheng Ning 1
; Qurat-Ul-Ain 1 ; Sahraoui Dhelim 3
1 School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing 100083, China;
2 Department of Computer Science, Blekinge Institute of Technology, 37179 Karlskrona, Sweden
3 School of Computing, Dublin City University, Dublin 9, Ireland