The potential of the metaverse in the field of education is an area of increasing interest, with many researchers exploring the space to increase the ease and efficacy of student education while reducing time and labor requirements to deliver effective teaching. However, there has been little work into the systematic and technological aspects of delivering education through the metaverse. To fill this gap, we propose a metaverse education system that takes good advantages of virtual reality and Web3 blockchain techologies to create a social learning environment. With this added emphasis on social aspects, learners are able to socialize and engage in collaborative efforts to improve their own knowledge. Using blockchain technology, the system can also help to ensure security and transparency while also keeping progression and grading fair for all participating students.
Education is fundamental for the growth and advancement of society because it helps all people understand new concepts, ideas, and methodologies to better the world. Understanding how people learn to offer an education system that achieves effective learning for all people has always been challenging. While some similarities exist, most students have significantly different preferred approaches to learning new concepts. For example, many of them prefer guided learning approaches to self-driven discovery learning . Ideally, all education would be personalized to each individual student’s preferences. However, the wide range of learning styles and varying degrees of aptitude makes it hard for traditional teaching methods to be universally effective, especially when considering personalized learning approaches. Furthermore, the current one-size-fits-all approach to education presents a barrier to students who would succeed if given personalized coaching .
To tackle this challenge, early research efforts have been devoted to intelligent tutoring systems (ITSs), where computational intelligence methods are used to mimic human tutors. As stated in a recent survey , the long history of productive research of ITSs has resulted in successful applications in education , military training , and healthcare , with even more work still ongoing. Early ITSs are often described as “homework helpers”, where a set of generalized or specific hints is provided upon a learner’s request . If a student were puzzled with a problem and failed to phrase a meaningful question, older ITSs might offer irrelevant or incorrect guidance that harms the student more than helps. With this in mind, ITSs continue to improve their mathematical student models through sensor informatics and machine learning. Rather than requiring students to ask relevant questions, modern ITSs monitor students’ behaviors in their learning and identify their individual needs for support . However, modern ITSs still have issues engaging students and providing interesting lessons. Furthermore, sensor informatics is a limited approach since many applications will not allow for the easy use of complex external sensors.
A second line of work that aims to overcome these shortcomings is to exploit the strengths of ITSs and increase student engagement through gamification. So called adaptive serious games use the principles of gamification to present educational concepts in an enjoyable and engaging setting. In other words, students can be distracted by game playing to the point where they do not recognize that they are learning. By adding intelligent or adaptive support, these games can be fully self-contained, providing lessons without the need for instructor intervention. While providing such personalized serious games is important with much potential benefits , many challenges still exist in the area. Although games provide a great environment to support contextualized knowledge construction, the requirement of self-directed and self-regulated learning on students makes it difficult to maximize a game's potential. While there is a wide range of data available in games for developers and researchers to analyze student performance and game effectiveness, physical learning data is still sparse as non-invasive physical sensors are often challenging to implement. Without necessary data, it is impossible to take good advantages of the power of data mining and artificial intelligence to build accurate and precise student/player models. Furthermore, many adaptive serious games offer one-player experiences, which do not consider the benefits of more social and group learning. However, recent technological advancement has made it possible for learning to occur anywhere and anytime. Any new and effective systems and platforms must consider that learning is no longer confined to classrooms, and must be able to capture learner information in any possible setting.
Metaverse, considered as the next generation of social connection , presents one potential solution to the aforementioned challenge in education. By extending physical learning through virtual and augmented technologies, physical education can be seamlessly integrated with virtual learning. By combining virtual reality learning with physical learning, an educational social space can be constructed where students are able to interact and socialize with peers while learning. Additionally, the flexible and configurable nature of virtual spaces makes it possible to tailor a wide range of lessons and educational approaches including personalized support. However, Metaverse education is still an emerging topic, with few efforts made to develop deep systematic approaches to this type of education system. Our prior work attempted to provide a systematic model for Metaverse education from the perspective of non-player characters (NPCs) that tutor students . In that work, we did not consider other types of NPCs that would learn alongside the students. In other words, we did not consider the benefits that social interaction these learner NPCs would bring to an educational system. And though these aspects are beneficial to consider, they also raise numerous issues with system security and safety. This paper aims to address the challenges and make the following contributions:
Extended from our prior work, this paper proposes MetaEdu, a novel framework that integrates both artificial intelligence (AI) and Web3 technologies through the ACP method (Artificial societies, Computational experiments, and Parallel execution) for effective Metaverse learning. MetaEdu considers that learning can occur everywhere, both within and outside of a standard classroom, including social interactions via extracurricular activities such as study groups. The three developmental phases of MetaEdu are then defined, and their relations are elaborated to show the progression and symbiosis of virtual and physical learning.
A detailed architecture of MetaEdu is then developed and analyzed, showing how the key technologies are applied to design various types of NPCs in the virtual space with the aim to optimize physical learning. In particular, blockchain technology is used to ensure the security, transparency, and fairness of shared social connection, while AI is deployed to provide students with an adaptive educational experience as they interact with MetaEdu.
The rest of the paper is organized as follows: Sect. 2 provides a review of relevant technologies that inspired the proposed system, and a discussion of outstanding issues with existing research. Section 3 presents the definition of MetaEdu, with the emphasis on its three developmental phases. Section 4 discusses challenges that could arise when moving forward toward the implementation stage of such a system, followed by our conclusions in Section 5.
2 Related work
2.1 ITS and serious games
ITSs have made great strides in recent years , sharing responsibility with instructors for estimating student knowledge and providing coaching and tutoring. Their effectiveness has been demonstrated in various fields of education, such as computer programming , language learning , dynamic system modelling , mathematics , and more general-purpose e-learning approaches . By providing students with more personalized education, ITSs aim to improve the efficacy of education while simultaneously reducing the strain on instructors’ limited time and resources. With an ITS, students can receive timely and personalized feedback on their learning without instructor intervention.
Among the various successful implementations, there are many AI methods that have been applied to map student data or performance into actionable system decisions. Methods like reinforcement learning  and genetic algorithms  allow AI systems to learn and adapt to new data. Other methods like Bayesian approaches  and fuzzy logic  allow experts to define their own logical behavior for AI tutors.
Beyond methods that focus solely on the AI side of ITSs, there has also been extensive developments in data mining , big data , and multimodal learning analytics  for educational approaches, with both areas showing promise for integration with more advanced AI methods. Methods like generative adversarial networks , unsupervised learning , and clustering  can work with student data to spot trends and make predictions that in turn can be used by AI methods to provide appropriate support.
A prominent field that extends the capabilities of ITSs is serious games, which are games made for education or training purposes. Serious games can be integrated with ITSs to both increase student engagement and to create a learning environment that focuses more on problem-solving. Principles of gamification [25, 26] are often applied to increase the educational merit and engagement of the system. And as such, serious games often focus on providing more immersive and exciting lessons compared to a standard ITS or classroom education. Beyond that, many of the technologies and systems established in the field of ITSs are also applicable within serious games such as reinforcement learning, supervised learning methods, and fuzzy logic .
As stated earlier, technological advances have made it easier to connect globally, resulting in vibrant networks of learners and content around the world. Learning communities are inevitably expanded beyond the boundaries of the classroom. However, both ITSs and serious games are primarily used in a traditional classroom setting or for one-on-one tutoring, despite their successful research and educational merit. Thus, there is a crucial need to bring ITSs and serious games into new development to address the emerging theme of “learning without borders” and many social situations where education is present.
The idea of the Metaverse has taken off in recent years with many researchers now exploring the possibilities and technologies of a shared virtual social space for work, school, and fun. The level of social connection, mobility, and collaboration in Metaverse presents great value to education, especially when considering the theme of “learning without borders”. Metaverse promotes deeper learning by naturally bringing learning into new contexts and allowing socialization for deep group collaboration . Gu et al. , for example, proposed using a metaverse and deep reinforcement learning to improve emergency evacuations, with a training system to help evacuees learn and predict efficient routes with a great improvement over traditional approaches . Artificial intelligence (AI) also plays very important roles in Metaverse to ensure proper arbitration, simulations, and decision-making . The involvement of AI in Metaverse makes it possible for data analytics that help better estimate learner knowledge for personalization. Similarly, blockchain technology can be fused into Metaverse, bringing education to a different level [31, 32].
Despite these prominent features of Metaverse for education, the research is still in its infancy. Besides heated discussions on its benefits and potential applications , there are very few technological developments. The design of virtual classroom with commercial-grade software and hardware is presented by Shen et al. to allow for a seamless connection between physical and virtual learning environments . Hare and Tang focused their efforts on building a virtual learning environment and designing AI-enabled tutor NPCs to offer guided learning . A case study of a consortium university in Korea for Metaverse education is presented in . Despite all these works, there is still a need for formal, systematic methods to guide the development of Metaverse education and, particularly, the integration of physical and virtual worlds to achieve optimal learning.
2.3 Parallel intelligent systems
With the advancement of system science and computer simulations, ACP (Artificial Systems, Computational Experiments, and Parallel Execution) methods were formally proposed by Fei-Yue Wang  to achieve Parallel Intelligence. ACP methods introduce a circular feedback mechanism to guide the operations of parallel intelligent systems - the integration of an artificial system with a real system. While the artificial system mirrors the actual system, computational experiments provide a unique way of testing models and algorithms in the virtual system that might be difficult or even impossible to conduct in the physical system. The optimal schema validated in the virtual system then has to act on the real system through parallel execution, including virtual-real interactions, double-feedback, and double closed-loop between the virtual and physical spaces.
In recent years, the ACP method has been widely applied to many domains. Ren et al. successfully used it to design a parallel vehicular crowd sensing (VCS) system . In particular, various computational experiments considering human and social factors were conducted, evaluated, and shared with the real VCS system to improve its efficiency and robustness. Similar studies can be found in transportation systems , healthcare , education , and image encryption .
Given these recent developments using the ACP methods and parallel intelligent systems, it can be said that there are many commonaltiies between parallel systems and metaverses. In particular, they both share the same challenges when dealing with complex systems. For example, there are many variables involved in operations of a complex system including many unknown latent variables. Understanding these variables is key to characterizing the complex system for any control and management application. However, such studies in the real world might be very costly or even impossible due to financial, legal, or institutional constraints. In this case, the ACP approach offers a viable solution. The successful application of ACP in other domains should be adopted for the design of Metaverse. Following this line of thinking, the proposed system focuses on applying an ACP approach to metaverse education to create MetaEdu.
It is clear that Metaverse has the potential to make education more flexible, interactive, and effective with equal learning accessibility. The more opportunities Metaverse present, the more complex learning systems become, and the more challenges have to be dealt with. Taking this into consideration, we propose a system called MetaEdu which aims to build a virtual learning world that starts from mirroring the physical world but goes far beyond it. MetaEdu is built to store users’ learning trajectories and knowledge trees irreversibly on the blockchain and establish a safe, fair, and open circle with credible data through partial disclosure. Unlike current virtual reality education, MetaEdu is also able to protect user privacy while keeping user information up-to-date through Web3, in addition to meeting the requirements of social interaction in educational conditions.
MetaEdu refers to a virtual-reality learning system based on metaverse technologies and features. It aims to generate a virtual clone of real-world learning environments and extend it to make the learning process more immersive for users. In addition to this, MetaEdu includes a blockchain technology-based Web3 reserve system that tightly integrates the virtual world with the physical world in terms of the learning system, social system, and identity system, and allows each user to produce specific content and edit the virtual world through their avatars. MetaEdu consists of three parts: the physical learning system for the world, Web3, and the virtual learning system.
The human world is the physical world of humans (teachers, students, etc.) who can communicate with each other and perform learning activities. The physical learning system aims to enable learning in the physical world, and therefore, it contains devices/hardware, systems, communication, and computing with educational applications. For example, books, personal communication devices, cloud computing devices, storage devices, management systems, and campus or social environments. The virtual learning system is a simulated system that can perform all learning operations in the physical world through artificial intelligence technology. It can also run and generate algorithms or systems designed as physical learning systems and store the results on Web3. In addition, its AI can interact with avatars of users in the human world through interactive devices. In contrast, users or robots in the human world can manipulate elements in the virtual learning system through Web3 to achieve MetaEdu’s integration of physical and virtual worlds.
The development of MetaEdu consists of three phases: clone, expansion, and fusion of surreality. The detailed development of the proposed system is given in Fig. 1.
The cloning phase refers to the mirroring process from the physical learning system to the virtual learning system. To give users a learning experience consistent with reality, the virtual world will have different scenarios that correspond to the physical world. For example, a classroom, library, and study room all located in the virtual space. These virtual scenarios must have the exact same elements and attributes as the physical world to encourage the same behaviors that users would perform in a physical learning environment. The end goal of the cloning phase is to allow users to experience a more convenient, efficient, and familiar virtual learning experience.
The expansion phase focuses on further developing and extending the framework created in the first phase. The main manifestation of this phase of work is that the virtual learning system will be improved and extended. At this stage, the virtual world as a mirror of the physical will be expanded with more scenarios and functions than the physical. For example, virtual classrooms that are easier to access with free technology experiments. In addition, virtual worlds are no longer just a mapping, but instead offer a way for students to self-improve beyond the limits of the physical world. Users participate in virtual worlds by logging into them to generate an avatar. Under the control of parallel strategies, the user’s behavior not only changes objects in the virtual world, but also generates impacts on the user experience in reality. Additionally, since the framework has already been built, the extended content of the virtual learning system will have a lower development cost with greater complexity and possibilities than the physical learning system. At the same time, however, security and privacy are critical factors to consider when transitioning to a virtual education system, including:
Cybersecurity threats: The teaching and learning resources of a virtual education system originate from the web and therefore may be vulnerable to cybersecurity threats such as hacking, malware, and phishing attacks.
Student safety: Virtual education systems may also pose greater risks to student safety, such as the possibility of cyberbullying or exposure to inappropriate content.
Data privacy: Virtual education systems often involve the collection and storage of student data, and online data storage may raise concerns about data privacy. It is therefore of utmost importance to ensure that student data is properly protected and collection and use of data is as transparent as possible.
The last phase is to deploy a multi-faceted interactive virtual reality system based on blockchain technology. In order to address the security and privacy issues raised in the second stage, the main goal is to ensure security, transparency, immutability, decentralization, and efficiency of information transmission between all participating parties. For these specific goals, blockchain technology offers a good solution. It is a decentralized and distributed technology that allows behavior and data to be securely recorded and verified without the need for a central authority. In MetaEdu, the physical system collects the user’s data and constantly updates a student model on the blockchain. This model can then be retrieved directly from the blockchain each time an educator or AI system calls for relevant content. Valid training results that need to be saved will also be uploaded to the blockchain to reduce storage risk.
Correspondingly, this new framework solves the problems of the original virtual world through the following aspects:
Security - Because blockchain is decentralized and distributed, it is more secure than traditional databases stored in a single location. This makes it more difficult for hackers to make unwanted changes to user information and records stored on the blockchain.
Transparency - Blockchain is a transparent system, which means that all learning records and the non-encrypted data stored on them are visible to anyone who has access to the network. This can help increase trust in the system.
Immutability - Once learning data has been added to the blockchain, it cannot be changed or deleted. This ensures that the information stored on the blockchain is accurate and cannot be tampered with.
Efficiency - Using blockchain to store user learning information has the potential to be more efficient than traditional databases because it eliminates the need for a middleman and can automate certain processes.
The three stages stated above also represent trends in human learning styles, so the systematic structure of the third stage will be explained in detail in the architecture.
The architecture of MetaEdu is shown in Fig. 2. As described in the previous chapter, MetaEdu is built on two worlds: the physical world and the virtual world. In MetaEdu, the two worlds interact and synchronize information through Web3-based on-chain connections to allow for independence and mutual feedback.
3.3.1 Physical world system
The physical world consists of three parts: Information Collection, Communication Computation and Storage, and Management and Control.
Information Collection (IC): The IC system handles all in-boundary and over-boundary transmission. The in-boundary transmission will include users’ information entry in the off-chain Internet, while over-boundary transmission covers the over-bound user information authentication, the over-bound update of the knowledge system framework, and sensor data such as voice recordings, gestures, expressions, heartbeat data, gaze tracking, or any other data collected when the user participates in MetaEdu.
Communication, Computation, and Storage (CCS): The CCS system is a system that enables the exchange of information, the processing of data, and the storage of data. The communication component of the system allows for the transmission of information between devices or systems through the internet. The computation component allows for the computational processing of data. The storage component allows for the preservation of data through the use of storage devices. Together, these three components enable the exchange, processing, and storage of information, allowing for efficient communication, data analysis, and data management.
Management and Control Center: The physical world management system and control system involves collaboration between teachers, school administrators, and other stakeholders in order to create a positive and effective learning environment for students. It also involves the combination of technological tools and pedagogical strategies online, as well as effective communication and collaboration between instructors, students, and other stakeholders. In particular, this system is also responsible for communicating with IC and CCS systems in our MetaEdu cycle, so as to complete on-chain user authentication, information upload, and knowledge framework update.
For the MetaEdu ecological cycle, the physical world system needs to rely on these three components for synchronization and feedback with virtual system:
IC systems to collect user authentication and feedback, update user learning status, and improve the on-chain model.
The CSS system to ensure user communication, collect and back up knowledge frameworks, and maintain efficient up-link communication. CSS is also responsible for outputting in-chain/ off-chain information to users.
The Management and Control Center to monitor and maintain the flow within the loop, using the best educational strategies to ensure that users learn easily and efficiently.
In relation to the blockchain, the chain stores not only the knowledge framework updated and kept by CCS, but also all the data of offline users, including login authentication data, interaction records and users’ knowledge records. In particular, due to blockchain irreversibility and on-chain publicness, MetaEdu can help users create on-chain knowledge trees with cascading updates to ensure fair and valid certification through group public scoring. Because of this, blockchain is a key technology that allows MetaEdu to operate more openly, fairly, securely, and efficiently.
3.3.2 Web3 system
Web3 refers to the next generation of the World Wide Web built on top of decentralized technologies such as blockchain. Web3 technologies are designed to allow users to interact with decentralized applications (dApps) and to take advantage of the security and transparency offered by blockchain. Blockchain in this case functions as a decentralized method of securely storing data and recording transactions. It consists of a network of computers that work together to validate and record transactions, which are then added to a chain of blocks that form a permanent record. Currently, blockchain is used for a variety of purposes, including the creation of digital currencies, the facilitation of financial transactions, and the storage and access of information which MetaEdu takes advantage of.
Blockchain in MetaEdu consists of 5 layers, as shown in Fig. 3:
Hardware/ Infrastructure layer: The hardware layer refers to the network of computers contributing to the blockchain’s computing power forms. A node is a computer or a network of computers that decrypt transactions.
Data storage layer: This layer is responsible for storing the data that is recorded on the blockchain. The data storage layer might use a variety of data structures, such as linked lists or hash tables to efficiently store and retrieve the data.
Network layer: This layer refers to the protocols that are used to connect the nodes in the network and enable them to communicate with each other.
Consensus layer: This layer is responsible for ensuring that all nodes in the network reach consensus on the state of the blockchain. It uses various algorithms and protocols to ensure that all nodes agree on the transactions that are included in the blockchain.
Application layer: This is the highest layer of the blockchain, and it refers to the applications and services that are built on top of the blockchain. These applications might include decentralized applications (dApps) and other services that allow users to interact with the blockchain and use its features.
In this layer structure, the primary function of the blockchain is to store and access information, and the various layers of the blockchain are structured in a way that enables this function to be performed efficiently and securely.
Users and virtual systems could access data on blockchain as shown in Fig. 4:
One of the smart contracts based on parallel intelligence can facilitate social interaction or interaction with other smart contracts; on the training model provided by the virtual system, contracts can be designed to allow testing and experimentation with different inputs or scenarios. Primarily, contracts are designed to allow the input of different variables or parameters and provide outputs based on these inputs.
It is worth noting, however, that the execution of smart contracts based on parallel intelligence is usually facilitated through the use of virtual machines, requiring consideration of the underlying blockchain platform as well as the capabilities and limitations of the smart contract. While parallel execution can be used in the contract itself, off-chain computation can also be used, or sharding can be used on the blockchain platform to improve overall efficiency and capacity.
3.3.3 Virtual world system
The virtual world system is a mirror and extension of the physical world that offers users a platform for personalized learning and communication. With AI-enabled non-player characters (NPCs), it can build a virtual learning system that revolves around the user’s physical world and their digital avatar, continuously optimizing learning methods and improving efficiency. The system is divided into two main parts, learner NPCs and tutor NPCs.
Learner NPCs, which act as peers to users, and can be either skilled learners or apprentice learners.
Skilled learner NPCs in MetaEdu exist to create more challenging and dynamic gameplay experiences for users. These NPCs exist to act as challenging opponents for users that react to user strategies in competitive situations to try to outperform users.
Apprentice learner NPCs in MetaEdu exist to ”learn” at a slower pace than users and skilled learner NPCs. Unlike skilled learner NPCs which exist to compete with users, apprentice NPCs instead offer users an opportunity to teach others. They act as peers to users to help them accomplish goals and help them achieve a deeper education through teaching others.
Skilled learner NPCs and Apprentice learner NPCs will store and share learning experiences through the blockchain while accessing information and data to learn and make decisions based on that information and data. They can also use natural language processing and AI methods to communicate and interact with students in meaningful ways. Behind the scenes, both types of NPC behaviors can be adjusted to ensure that students receive appropriate competition or guidance from both competitive and collaborative NPCs. And while these NPCs may have conflicting goals, educational scenarios can be tailored carefully to students to ensure that NPCs only act when it is appropriate for collaboration or competition.
Unlike learner NPCs which function as peers, Tutor NPCs in MetaEdu are meant to create more effective educational experiences. Tutor NPCs can be used to present information and explanations, provide examples and practice exercises, and offer feedback and reinforcement to help students improve their understanding and performance. This could be particularly useful in online or distance learning environments, where students may not have access to a human instructor.
Tutor NPCs access information about learning frameworks and student users via the blockchain. Using machine learning algorithms to analyze data about the student’s performance and learning progress, tutor NPCs adjust the learning experience accordingly. The NPC will provide more or less challenging material based on the student’s performance, or may focus on specific areas where the student is struggling. This can help ensure that the learning experience is tailored to the student’s needs and abilities, and can help them progress more quickly and effectively. Some additional details and possible methods of NPCs were addressed in our prior work .
To provide students with an adaptive learning experience in the virtual world, we use the model shown in Fig. 5. This computational experiment model is built to be highly controllable, easy to apply, and easily reproduced. In Fig. 5, the inputs \(F_a, F_b,..., F_n\) are factors collected by the system. For example, the system might collect score on an exam, time taken to complete the exam, and gaze tracking data on which question the student looked at longest. The optimization model is then trained on this data to estimate student performance and select what guidance those students require. While specific methods to translate student data into knowledge models are beyond the scope of this paper and left up to implementation, the system may, for example, score the user in several categories using clustering methods. It would then select a hint from a database of hints, or generate a paragraph of useful information using a natural language model. In addition to providing support to the learner in the physical world, student models can also feedback to help improve the behavior of NPCs and make them more realistic (for learner NPCs) or more effective (for tutor NPCs). With this parallel approach, the goal is to improve system performance on multiple fronts while helping the user learn.
While MetaEdu presents a good framework for a new way of education, there are many challenges ahead.
Security: since MetaEdu is a very complex system involving multiple smaller systems, it has many privacy and security issues. On a system level, the virtual world is a clone of the physical world, which naturally contains geographic information; the virtual learning world could also contain sensitive knowledge that needs to be taught, such as proprietary information from industries or countries. From the user level in MetaEdu, human users interact with in the digital world through virtual reality devices, and the personal and activity data collected by the devices are stored in the MetaEdu blockchain. The loss or leakage of information during the transmission process could cause huge losses to the user or related users. At the same time, a large amount of user information and knowledge models are stored on the chain, and it is very important to protect their security and integrity. However, since the number of MetaEdu users is huge and the knowledge system is constantly expanding, protecting their privacy and security is also an important challenge for MetaEdu.
Intelligence: to achieve the goal of introducing teaching and learning into both the physical and virtual worlds, MetaEdu relies on artificial intelligence (AI) to build various non-player characters (NPCs) that present diverse challenges in terms of intelligence requirements. On the one hand, since NPCs in virtual worlds have changing goals and environments, an AI model that can continuously learn and update itself is required. On the other hand, multiple training models exist in the system from top to bottom, and they need to be trained on all of the collected data. This information has considerable complexity and dimensionality, putting tremendous pressure and difficulty on the training. Therefore, adding a layer of trainers that can dynamically filter and update the training data set is a possible solution that would ensure smoother operation of the completed MetaEdu system.
Computation: as we mentioned in the previous point, as the number of users increases and the knowledge architecture is updated, a stable and efficient system ecology is necessary. So, without degrading the user experience, MetaEdu needs a system that can provide great computing power. It must have a large amount of storage space, fast computing power, and at the same time be responsible for managing system processes while maintaining stable operation of the system within a manageable latency.
In order to break the boundaries of the traditional education model and push education to a higher platform, we apply the concept of Metaverse to education and propose MetaEdu. MetaEdu is an educational system that enables learning and communication simultaneously in the physical and virtual worlds, greatly improving learning efficiency while enabling secure, seamless connections and interactions between users. The development stages of MetaEdu include cloning, extending, and surreality fusing to put together the physical and virtual components and the blockchain technology necessary to enable the completed system. By connecting both through the blockchain, the MetaEdu framework allows for safe and secure collection and storage of user data to enable powerful AI techniques, all with the end goal of enhancing student learning. Using the ideas outlined in this paper, we hope to inspire future researchers to create and apply MetaEdu to offer more effective and efficient education to students around the world.
Data sharing is not applicable to this article as no data were generated or analysed during the study.
Hattie J. Visible learning: a synthesis of over 800 meta-analyses relating to achievement. London: Routledge; 2008. https://doi.org/10.4324/9780203887332.
Tang Y, Franzwa C, Bielefeldt T, Jahan K, Saeedi-Hosseiny MS, Lamb N, Sun S. Sustain City: effective serious game design in promoting science and engineering education. Hershey: IGI Global; 2023. p. 914–43. https://doi.org/10.4018/978-1-6684-7589-8.ch044.
Liang J, Hare R, Chang T, Xu F, Tang Y, Wang F-Y, Peng S, Lei M. Student modeling and analysis in adaptive instructional systems. IEEE Access. 2022;10:59359–72. https://doi.org/10.1109/ACCESS.2022.3178744.
Jing S, Tang Y, Liu X, Gong X, Cui W, Liang J. A parallel education based intelligent tutoring systems framework. In: 2020 IEEE International Conference on Networking, Sensing and Control (ICNSC), pp. 2020:1–6. https://doi.org/10.1109/ICNSC48988.2020.9238052.
Lafond D, DuCharme MB, Rioux F, Tremblay S, Rathbun B, Jarmasz J. Training systems thinking and adaptability for complex decision making in defence and security. In: 2012 IEEE International Multi-Disciplinary Conference on Cognitive Methods in Situation Awareness and Decision Support, pp. 2012:51–58. https://doi.org/10.1109/CogSIMA.2012.6188408.
Koutsojannis C, Prentzas J, Hatzilygeroudis I. A web-based intelligent tutoring system teaching nursing students fundamental aspects of biomedical technology. In: 2001 Conference Proceedings of the 23rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol. 4, 2001:4024–40274 . https://doi.org/10.1109/IEMBS.2001.1019728.
Chad H. Intelligent tutoring systems : prospects for guided practice and efficient learning. 2006.
Hare R, Tang Y. Player modelling and adaptation methods within adaptive serious games. In: 2021 International Conference on Cyber-Physical Social Intelligence (ICCSI), 2021:1–6. https://doi.org/10.1109/ICCSI53130.2021.9736213.
Hwang G-J, Chien S-Y. Definition, roles, and potential research issues of the metaverse in education: an artificial intelligence perspective. Comput Educ: Artif Intell. 2022;3: 100082. https://doi.org/10.1016/j.caeai.2022.100082.
Hare R, Tang Y. Hierarchical deep reinforcement learning with experience sharing for metaverse in education. IEEE Trans Syst Man Cybern Syst. 2022. https://doi.org/10.1109/TSMC.2022.3227919.
Gobert JD, Sao Pedro MA, Li H, Lott C. Intelligent tutoring systems: a history and an example of an its for science. In: Tierney RJ, Rizvi F, Ercikan K, editors. International encyclopedia of education. 4th ed. Oxford: Elsevier; 2023. p. 460–70. https://doi.org/10.1016/B978-0-12-818630-5.10058-2.
Hooshyar D, Ahmad RB, Yousefi M, Fathi M, Abdollahi A, Horng S-J, Lim H. A solution-based intelligent tutoring system integrated with an online game-based formative assessment: development and evaluation. Educ Technol Res Dev. 2016;64(4):787–808. https://doi.org/10.1007/s11423-016-9433-x.
Vogt P, van den Berghe R, de Haas M, Hoffman L, Kanero J, Mamus E, Montanier J-M, Oranç C, Oudgenoeg-Paz O, García DH, Papadopoulos F, Schodde T, Verhagen J, Wallbridgell CD, Willemsen B, de Wit J, Belpaeme T, Göksun T, Kopp S, Krahmer E, Küntay AC, Leseman P, Pandey AK. Second language tutoring using social robots: A large-scale study. In: 2019 14th ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2019:497–505. https://doi.org/10.1109/HRI.2019.8673077.
Zhang B, Jia J. Evaluating an intelligent tutoring system for personalized math teaching. In: 2017 International Symposium on Educational Technology (ISET), 2017:126–130. https://doi.org/10.1109/ISET.2017.37.
Ennouamani S, Mahani Z. An overview of adaptive e-learning systems. In: 2017 Eighth International Conference on Intelligent Computing and Information Systems (ICICIS), 2017:342–347. https://doi.org/10.1109/INTELCIS.2017.8260060.
Georgila K, Core MG, Nye BD, Karumbaiah S, Auerbach D, Ram M. Using reinforcement learning to optimize the policies of an intelligent tutoring system for interpersonal skills training. In: Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems, 2019:737–745.
Stein G, Gonzalez AJ, Barham C. Machines that learn and teach seamlessly. IEEE Trans Learn Technol. 2013;6(4):389–402. https://doi.org/10.1109/TLT.2013.32.
Hooshyar D, Ahmad RB, Wang M, Yousefi M, Fathi M, Lim H. Development and evaluation of a game-based bayesian intelligent tutoring system for teaching programming. J Educ Comput Res. 2018;56(6):775–801. https://doi.org/10.1177/0735633117731872.
Papadimitriou S, Chrysafiadi K, Virvou M. Fuzzeg: fuzzy logic for adaptive scenarios in an educational adventure game. Multimed Tools Appl. 2019;78(22):32023–53. https://doi.org/10.1007/s11042-019-07955-w.
Baker RS. Educational data mining: an advance for intelligent systems in education. IEEE Intell Syst. 2014;29(3):78–82. https://doi.org/10.1109/MIS.2014.42.
Baneres D, Caballé S, Clarisó R. Towards a learning analytics support for intelligent tutoring systems on mooc platforms. In: 2016 10th International Conference on Complex, Intelligent, and Software Intensive Systems (CISIS), 2016:103–110. https://doi.org/10.1109/CISIS.2016.48.
Chui KT, Liu RW, Zhao M, De Pablos PO. Predicting students’ performance with school and family tutoring using generative adversarial network-based deep support vector machine. IEEE Access. 2020;8:86745–52. https://doi.org/10.1109/ACCESS.2020.2992869.
Hershcovits H, Vilenchik D, Gal K. Modeling engagement in self-directed learning systems using principal component analysis. IEEE Trans Learn Technol. 2020;13(1):164–71. https://doi.org/10.1109/TLT.2019.2922902.
Bunić D, Jugo I, Kovačić B. Analysis of clustering algorithms for group discovery in a web-based intelligent tutoring system. In: 2019 42nd International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), 2019:759–765. https://doi.org/10.23919/MIPRO.2019.8756951.
Bakhanova E, Garcia JA, Raffe WL, Voinov A. Targeting social learning and engagement: what serious games and gamification can offer to participatory modeling. Environ Model Softw. 2020;134: 104846. https://doi.org/10.1016/j.envsoft.2020.104846.
Fleming T, Sutcliffe K, Lucassen M, Pine R, Donkin L. 10.05-Serious games and gamification in clinical psychology. In: Asmundson GJG, editor. Comprehensive clinical psychology. 2nd ed. Oxford: Elsevier; 2022. p. 77–90. https://doi.org/10.1016/B978-0-12-818697-8.00011-X.
Esfahlani SS, Cirstea S, Sanaei A, Wilson G. An adaptive self-organizing fuzzy logic controller in a serious game for motor impairment rehabilitation. In: 2017 IEEE 26th International Symposium on Industrial Electronics (ISIE), 2017:1311–1318. https://doi.org/10.1109/ISIE.2017.8001435.
Zhong J, Zheng Y. Empowering future education: Learning in the edu-metaverse. In: 2022 International Symposium on Educational Technology (ISET), 2022:292–295. https://doi.org/10.1109/ISET55194.2022.00068.
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. https://doi.org/10.1109/TSMC.2022.3231299.
Hwang G-J, Chien S-Y. Definition, roles, and potential research issues of the metaverse in education: an artificial intelligence perspective. Comput Edu: Artif Intell. 2022;3: 100082. https://doi.org/10.1016/j.caeai.2022.100082.
Tlili A, Huang R, Shehata B, Liu D, Zhao J, Metwally AHS, Wang H, Denden M, Bozkurt A, Lee L-H, et al. Is metaverse in education a blessing or a curse: a combined content and bibliometric analysis. Smart Learn Environ. 2022;9(1):1–31. https://doi.org/10.1186/s40561-022-00205-x.
Li J, Lan M, Tang Y, Chen S, Wang F-Y, Wei W. A blockchain-based educational digital assets management system. IFAC-PapersOnLine. 3rd IFAC Workshop on Cyber-Physical Human Systems CPHS 2020. 2020;53(5):47–52. https://doi.org/10.1016/j.ifacol.2021.04.082.
Wang M, Yu H, Bell Z, Chu X. Constructing an edu-metaverse ecosystem: a new and innovative framework. IEEE Trans Learn Technol. 2022;15(6):685–96. https://doi.org/10.1109/TLT.2022.3210828.
Shen T, Huang S-S, Li D, Lu Z, Wang F-Y, Huang H. Virtualclassroom: a lecturer-centered consumer-grade immersive teaching system in cyber-physical-social space. IEEE Trans Syst Man Cybern: Syst. 2022. https://doi.org/10.1109/TSMC.2022.3228270.
Jeon JH. A study on education utilizing metaverse for effective communication in a convergence subject. Int J Internet Broadcast Commun. 2021;13(4):129–34. https://doi.org/10.7236/IJIBC.2021.13.4.129.
Wang FYWP. Intelligent systems and technology for integrative and predictive medicine: an acp approach. ACM Trans Intell Syst Technol. 2013. https://doi.org/10.1145/2438653.2438667.
Ren Y, Jiang H, Feng X, Zhao Y, Liu R, Yu H. Acp-based modeling of the parallel vehicular crowd sensing system: framework, components and an application example. IEEE Trans Intell Veh. 2022. https://doi.org/10.1109/TIV.2022.3221927.
Shi H, Liu G, Zhang K, Zhou Z, Wang J. Marl sim2real transfer: Merging physical reality with digital virtuality in metaverse. IEEE Trans Syst Man and Cybern: Syst. 2022. https://doi.org/10.1109/TSMC.2022.3229213.
Aloqaily M, Elayan H, Guizani M. C-healthier: a cooperative health intelligent emergency response system for c-its. IEEE Trans Intell Transp Syst. 2022. https://doi.org/10.1109/TITS.2022.3141018.
Gao Y, Miao H, Chen J, Song B, Hu X, Wang W. Explosive cyber security threats during covid-19 pandemic and a novel tree-based broad learning system to overcome. IEEE Trans Intell Transp Syst. 2022. https://doi.org/10.1109/TITS.2022.3160182.
Li P, Sun Z, Situ Z, He M, Song T. Joint jpeg compression and encryption scheme based on order-8-16 block transform. IEEE Trans Intell Transp Syst. 2022. https://doi.org/10.1109/TITS.2022.3217304.
The authors declare no competing interests
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Cui, L., Zhu, C., Hare, R. et al. MetaEdu: a new framework for future education. Discov Artif Intell 3, 10 (2023). https://doi.org/10.1007/s44163-023-00053-9
- Metaverse learning
- Artificial intelligence
- Parallel Intelligence