UMAP 2013 Conference on User Modeling, Adaptation and Personalization

Last week I attended the UMAP conference in Rome, where I could watch several interesting presentations and talk to inspiring people. Of course, my priorities have been influenced by our projects ROLE, GALA, and Layers. Perhaps some of the UMAP papers can stimulate new research and development directions in this context.

Although learning represents just a small niche at the conference, this year both awarded papers came from this field. Especially the work presented by Jennifer Sabourin raised my interest, as it is related to Self-Regulated Learning (SRL). It is based on the finding that students with metacognitive (e.g. SRL) skills are more successful in learning, so they attempt to identify them early enough. This knowledge can be used later on for suitable training of these skills or adaptive scaffolding. For their predictions they are using highly structured game-based learning environments, but this finding can be perhaps used later on in more open-ended exploratory environments too. As they note, it will be important to measure the impact of adaptive scaffolding both in terms of learning performance and user engagement, as too much guidance can reduce the learner’s motivation, so it must be optimally balanced with the learner control. A right support of SRL strategies is a big challenge.

Monitoring physiological parameters can provide a valuable feedback for adaptive learning. In addition to other sensors, Brain Computer Interfaces became recently more affordable and less intrusive. Georgios Patsis used such more usable technology to measure attention level in the Tetris game and was able to achieve similar outcomes like those previously obtained with a 19 sensor EEG cap. This offers new opportunities to adapt games and learning processes to the mental status of the user.

Also in the Lifelong User Modelling workshop Judy Kay dealt with meta-cognitive scaffolding, especially how to support self monitoring of long term goals. This kind of reflection should be supported by Open Learner Modelling, where visualization plays the key role. Thus lifelong user model provides the foundation to support meta-cognitive scaffolding for self monitoring, reflection, and goal setting.

The visualization dimension was the primary focus of the User-Adaptive Visualization workshop, which is establishing a working group with the aim to coordinate this research and develop a taxonomy of adaptive visualization techniques. People interested in this area can join their LinkedIn group.

There were also other interesting contributions, like the paper by Dietmar Jannach on an analysis of recommender systems, the poster by Saskia Koldijk on unobtrusive monitoring of knowledge workers for stress self-regulation, or contributions from the ImREAL project. These and other ones can be found in the conference proceedings or in the extended proceedings.

Leave a Reply