AI Injected e-Learning

The areas of Artificial Intelligence (AI) and e-Learning from the Computer Science and Education domains respectively are not usually associated together, not because they are not compatible or complementary but due to a number of other non-technical reasons.

e-learning has come a long way and it is only thanks to previous versions and numerous evolutions of e-learning that we can propose new routes and design intelligent systems for future generations. This also enables us to appreciate and value the meaning of moving forward as we fully understand and acknowledge from where we are coming. A plethora of research studies have reported conflicting results over the years as some praise and applaud this medium, while others disapprove and critise e-learning in every possible way. The fact that e-learning itself is not regulated by a specific academic body and that best practices are subjective, divergent and too generic, renders the whole playing game fuzzy, confusing and incredibly frustrating to the learners. However, pedagogical trends and technological forces have shaped the history of e-learning and will continue to do so. How have these rubbed off onto each other? And how have they influenced the following generation of e-learning? What are the factors that will impinge on the future of online education? In this chapter a deeper examination and appreciation of these changes and developments over the years is presented in an effort to understand the inevitable evolution that occurred and how this affected and influenced the whole environment surrounding e-learning. These include the social implications, the pedagogical repercussions and the technological impacts that gave rise to different e-learning generations.

Social media took the world by storm and transformed the society and its multiple dimensions in more than one way. The extent of the shock waves that this phenomenon inevitably influenced the way people interacted with the web and with each other, as well as with all web applications and services provided online. E-learning evolved as it embraced the new Web 2.0 technologies in an attempt to enhance the delivery but at the same time to take full advantage as in the past of the latest cutting-edge technologies that were available. It has been argued in the previous chapter that this technological shift was no standard evolution but a major unconventional and progressive e-learning revolution that literally turned the tables around. In this chapter the full impact of this considerable technological contribution to the pedagogical and functional dynamics of e-learning will be brought into perspective as innovative techniques transpired from the evolution of Web 2.0 technologies that slowly but surely got integrated within online learning systems. These include MOOCs or Massive Online Open Courses, Crowdsourcing techniques, and Social Networks. The beauty about these technologies that resulted out of the latest technological evolution addressed particular e-learning concerns as e-learning had been emanating from the integration of a variety of incongruous emerging technologies that at the time assisted in improving the services provided by such systems. In the following sections the main e-learning issues will be discussed together with how emerging technologies can start addressing them.

Personalisation, user profiling and the use of machine learning techniques from the computer science arena fall under the umbrella of Artificial Intelligence or AI. Rather then going through all the technical details of machine learning and AI we will be looking into the conceptual application of such techniques, as well as the educational undertones of doing so. Personalisation features as a main component in this chapter due to its exceptional and remarkable property of improving a service or a product. We shall be looking into how such a widely employed technique in industry can be similarly applied to education that promises to alleviate and add-value to e-learning as we know them. The main concept behind such a technique is the capturing and representation of the specific user model or profile. This user representation is a living model that evolves over time and requires constant updating to ensure the profile realistically embodies the user or the learner in our case. As we shall investigate in the next sections the user profile is generally generated and trained using the user patterns and trends but also the interests, needs and choices that all indicate something specific about the user in isolation as well as in combination together. In another section we will also take an in-depth analysis of how user profiling can be optimised in the case of education in a similar attempt to encapsulate the specific and characteristic learner profile. We close this chapter with a look at recommender systems and how all the different parts mentioned above come together to the cause of enhancing education and the e-learning medium.

The initial steps towards the model to personalising e-learning through the injection of AI starts to take shape in this chapter as several of the factoring elements have been covered in the previous two chapters. These will form part of a personal learning environment that each individual learner or life-long learner would establish and assemble around oneself in an effort to create and avail of a sustainable educational system that has the learner at its centre. Personal learning environments or PLEs are ideal vessels to encapsulate all a learner requires due to their personalisation capabilities that truly empower the same learner. Morrison [1] identifies two essential components within a PLE as he depicts its anatomy as shown in Fig. 5.1 overleaf. Each of these components play an important role and need to be investigated individually to ensure that they are optimally setup and compatibly designed to generate the expected outcome, an intelligent personal learning environment. The Personal Learning Network (PLN) and the Personal Learning Portfolio (PLP) form part of the PLE and will be initially presented in the following sections as they bring together essential components from the previous chapters. These underlying technologies that source both of them will be justified in terms of their academic relevance, pedagogical effectiveness, and theoretical suitability. How will both these components take advantage of the latest technological developments and boost well established technologies in an attempt to enrich the learning experience? How will the different technologies compatibly come together to ensure the learner is not just at the centre of the PLE but also in full control of the medium? The personal learning network and the personal learning portfolio are intended to complement each other as they form part of the proposed model in the next chapter.

This chapter brings all the previous chapters together as they collectively and incrementally built up a crescendo to reach the highlight, namely injecting e-learning with AI to customise the education process. The proposed model makes us of all the techniques discussed in the previous chapters and endeavours to compatibly bring them together to create an intelligent personal learning environment. The evolution of e-learning was led imposed by the technology but this model proposes to conveniently employ numerous technologies and techniques to directly address specific e-learning issues. The next generation of online education is not dictated by technology but by the academic need to personalise learning together with the efficient automation offered by AI. The first e-learning issue addressed is that of isolation and Chap. 3 undertook this task with the ingenuity of crowdsourcing and the popularity of social networks. The connectivism learning theory has been associated with this phenomenon and this model makes good use of this first factor. Motivation is another e-learning issue that is addressed through the contributions from Chap. 4 as learner profiling and learning portfolios support student to be much more self-determined in their academic endeavour. The third and final issue tackles the issue of impersonality that e-learning is notoriously criticised, and Chap. 5 offers adaptive environments through the combination of a learning portfolio and supportive learning network. A truly intelligent personal learning environment backed and injected by AI techniques is being proposed as a compatible combination of all these technologies to enhance e-learning effectiveness as it leads online education to its future and the next e-learning generation. The rest of this chapter is organised as follows. The section that follows expands further the underlying rationale that led to the proposed model by analysing the contributions from the previous chapters. This is followed by the architectural setup of how these technologies come together within an online system to deliver a functional and intelligent PLE. The next section tackles all the implementation details that take place to accomplish and complete the architectural design presented before. Finally operational and pragmatic details of how the online PLE functions are covered in an effort to show how the AI-injected e-learning system will operate in reality.

The book draws to an end by looking ahead at potential future avenues in light of the proposed intelligent personal learning environment. Web technologies and AI techniques continue to evolve as e-learning systems continue to take full advantage of both to improve the delivery and the overall holistic experience. The employment of AI techniques in combination with other technologies moved away from the conventional trend of adopting the latest web technologies to embellish the e-learning environment and move to the next generation. The proposed intelligent learning environment had set objectives with specific issues to resolve and employed the different methodologies and practices within an original architectural setup that fulfils the pre-set e-learning needs. Will it be possible to pursue this trend whereby the e-learning needs dictate and prescribe what the technology should be like and impose what it should provide? On the other hand the same architectural setup introduced a novel concept of bringing together numerous technologies to achieve a common goal, personalised e-learning. Will future e-learning generations persist on this line of thought and take full advantage of multiple developments in numerous and diverse domains to collectively achieve a superior added-value outcome that could potentially shape the future of e-learning? This final chapter looks ahead at these possibilities and the potential of influencing future e-learning generations by reversing the way e-learning advocates reason and devise such futuristic environments. Which technological novelties will leave their impact on future e-learning setups? What exactly is the ideal e-learning scenario and which technologies or combination of technologies can pave the way forward?