WP5: Support and feedback during learning

Objectives
1. System generated recommendations on interactions based on automatic analysis of learner interactions within the learning environment.
2. System generated feedback on writing style and recommendations on selection of learning topics based on automatic analysis of learner produced discourse.

Description of work
This work package is concerned with support during learning processes. It researches the techniques for
analysis and recommendations and designs and creates the services for process support and feedback. Two specific areas for support and recommendation services are being identified. First, support based on analysis of interactions between learners (task 1) and, second, support based on the analysis of discourse produced by the learner (task 2). For task 1 - recommendations based on interaction analysis - the interactions of students in terms of interaction data (discussion forums and chats) and produced content are semi-automatically analysed with language technologies based on Natural Language Processing (NLP) and with Social Network Analysis (SNA). The technologies will be used to sort and archive discussion threads; to assemble overviews and indicators of the activities of a community as a whole and of the individual actors and to recommend peers. The results will be used to improve and/or increase the level of feedback to tutors, moderators, or to the learner as an individual or in their team context. In task 2 - recommendations based on assessing textual products - the learners’ comprehension of a domain will be examined through an analysis of their textual products. The analysis will provide recommendations on style, structure, content and coherence of the learners’ products. The recommendations, combined with guidelines, can be used by the tutor, the learner or a group. This task reuses some techniques using LSA that are well-known and well established in the literature, such as coherence assessment, and long text comparison. The work will have a combined focus on both modelling text comprehension and assessing students’ writing and will include research for the assessment of very short notes; the automatic extraction of questions from short notes, and the format of knowledge visualization that would be useful for collaborative learning.
Participants: UU, WUW, UPMF, PUB-NCIT, UNIMAN