Portal:Complex Systems Digital Campus/POEM

POEM Personalized Open Education for the Masses

Description
Aims at designing an open Social Intelligent ICT platform for a Lifelong Personalised Education Ecosystem that reconciles the scalability of MOOCs with Personalization. Personalization comprises active participation of the students in their own education; predictive strategies for the next step toward the preferred skill set, and knowledge; finally, preventive strategies for avoiding future difficulties, especially withdrawals.

In the envisioned ecosystem, learning can be seen as a journey among pedagogical resources that you have to see, to understand, to integrate, and to remember. This journey is guided by a multidimensional knowledge map where students go from one geo-localized location (the prior knowledge) to another as destination (the learning goals). Personalization can be reached by actively managing the user profiles in terms of prior knowledge and learning goals. A MOOC is participative if the learner defines herself the destination advised by the educational ecosystem. As the ecosystem records all paths on the knowledge map, it can be predictive by matching the actual learner profile to the recorded ones. Thus it can predict the issue of the current state (e.g., this course can be difficult to you, so pay attention) and next states (e.g., you will probably do this among what others did). Prediction is a basis for prevention (avoid paths to dead ends).

Like a human tutor of the ancient times, the educational ecosystem may not always know the absolutely best learning path, but it can offer trajectories that have been successful in the past.

The first aim of educational ecosystem is to provide a Social Intelligent ICT platform for a Long life Personalized Education Ecosystem The rapid evolution of economy and society in the age of globalization requires an also rapid evolution of competences and thus necessitates life-long education, training, and re-training. At the same time education must become more personalized, as the competences needed to master an increasingly sophisticated technological world continually diversify. The project aims at designing an open Social Intelligent ICT ecosystem for a Lifelong Personalised Education Ecosystem. In other words the ecosystem has to reconcile the strong development of MOOC with Personalized Education. Personalized education must have the following specific requirements as with the personal tutor in the ancient time: firstly, a strong personal involvement of the students to participate actively to their own education; secondly, predictive strategies for guessing the next step toward the best skill level in the preferred domain of knowledge for each student; finally, preventive strategies for avoiding future difficulties, especially withdrawals. We call such conjunction of participative, predictive, preventive and personalized education the 4P Education or more simply Personalized Education, if the emphasis is not on the preconditions but on the result. It can seem that massive education and personalized education are antagonist objectives. Let us first remark before continuing that, at the contrary, they are in synergy. Without a massive number of previous educational trajectories, the educational ecosystem will be not able to guess the best future for a new one. Best guessing supposes the data assimilation of a massive number of such trajectories. Thus, the participation of everybody to such educational ecosystem is the most desirable World for any human at any age to come back for actively and quickly learn something more and different from what he knows. But such synergy between massive and personalized education is only possible within a social intelligent ICT ecosystem. What follows describe the Lifelong Personalized Education Ecosystem as such social intelligent ICT. It will be tested on three partner’s organizations: OU, FGV and the University of Strasbourg.

Description
The educational ecosystem will involve the students in a lot of individual and collective educational activities for their mutual benefit: assessment, inter-tutorship, construction of dynamical Knowledge Maps. All these activities are in relation with the main functions of the educational system, e.g.:

A first function is to construct and visualize the dynamic Knowledge Maps of domain as well as, inside, the individual, the MOOC and the Curriculum trajectories. This function is useful for students when they have personalized choice to do and to professors for examining the coherence of their MOOC inside a Curriculum or the coherence between Curriculum. Conversely, crowdsourcing by students is useful for extracting a corpus of a new domain for constructing a new Knowledge map.

A second function is the inter-tutorship: each student has a tutor, still a student more advances in the same curriculum; he can ask him questions; when there is no response the tutor can forward the question to his own tutor and again and again.. until there is no tutor and the question arrive to a professor.

A second function is assessment that produces a series of success failure along the personalized trajectories of students. Different automatic assessment methods will be proposed, sophisticated MCQ, open responses to open questions[1], automatic correction of exercises[2]. But peer-to-peer will be also used either because they are known to be better either because they have to be compare with automatic assessment until there is no doubt about the best one. Peer-to-peer assessment is recognised to be of mutual benefit. Whatever is the assessment method to the tests, this method provides a skill-level order like Elo-points in chess or Tennis ranking. The skill-level of students and the skill-level required for tests are coevolving until the student at a skill-level are successful until the test at the same level.

The individual and crowdsourcing activities will be observed by the Educational Ecosystem. Online data assimilation of individual educational trajectories will allow to create an integrated individual profile (including a skill-level) as the best summary of the past[3] activities of the student. Similarly, data assimilation of crowdsourcing activities will allow to create an integrated social profile (including a reputation-level) by best summarizing the past. The reputation-level is increasing for a tutor whose the student has a rapidly increasing skill-level.

Description
An essential quality criterion of personalized education is the maximum of skill-level in case of success to the best preferred curriculum. It is obtained through a best individual educational trajectory. The main problem is the same as those of the ancient personal tutor: guessing for the student a very limited number of next steps for his best trajectory…keeping to the student the final guess of the next step.

It is conjecturing that, given an individual profile, the best next incremental step is determined in probability by the distribution of the choices of previous learners with the same profile. This conjecture is the Personalized Educational Man-Hill Problem, because of the similarity with ants collective behaviour.

The meaning of this conjecture is that the educational ecosystem plays exactly the same role as the ancient personal tutor. But it simply collects all the educational trajectories, categorizes them at present time through personal profile and use the empiric distribution of next steps observed for the same profile of previous trajectories. The most difficult operation is the categorization This conjecture hold when the number of educational trajectories tends to the infinity. But it is never the case because the dynamic Knowledge Map is never stationary, because the arrival of new hot topics, that can be provided by crowdsourcing of students! And the students can observe the scientific trajectories of these hot topics.. and even better observe how the old domains are splitting or slicing into new domains.

Description
The other essential quality criterion of personalized education is the minimum of failure toward the best preferred curriculum. The indicator to be predicted and checked is the withdrawal probability (and, secondary, the predicted distribution of probability of the withdrawal time). If the withdrawal probability increases above a threshold, the prevention starts by offering the learner some remediation training drills to prevent failure. If the withdrawal probability continues to increase, recurrent contact with a more tutor with higher reputation-level is provided. If the withdrawal probability is overcoming the next threshold, a discussion with a professor is offered. The prevention process follows the same kind of path as in preventive medicine.

Description
As already, the two main quality criteria of an individual performance is success/failure and the skill-level attained in case of success. A MOOC, a Curriculum or a Personalized Educational Ecosystem Performance is simply the mean of the individual performances. If a MOOC has bad performance, the psychophysical and neurophysiological measurements present in the project can diagnostic which parts obtained less attention of the students. The presence of clear criteria of performance is a crucial factor for a permanent amelioration of Curriculum and Personalized Educational Ecosystem. Such Personalised Educational Ecosystem is taking the advantages both of the connectionist MOOCs with its crowdsourcing activities and of traditional MOOCs with their automatic assessment. It will be clearly beyond the state of art in automatic assessment for MCQ, open response to open question and exercises. But, it will radically beyond the state of art for personalized education. If it is successful, it means that the Personalized Education can be attained in the next years.

Return to the Wikiversity Portal of the Complex Systems Digital Campus

Return to the Complex Systems Digital Campus main website