We are organizing a workshop! Together with some of my colleagues at the lab, we are organizing a workshop at the UMAP conference this year on the representation of context and its impact on user modeling. These are my thoughts.
How it started
One afternoon, one of my colleagues at the ADAPT lab, Jovan Jeromela, brought it up during lunch that our guide, Prof. Owen Conlan, wanted us to run a workshop. “It would be very useful for research experience,” he claimed and the researchers there ruminated on the prospect of having our own workshop. It seemed thrilling enough, so we set up a google calendar event to discuss what exactly we would have our workshop on.
The subject of the workshop is a topic of contention for our group. This is because of the group of people who decided to work on organizing this workshop, none of us work on even remotely the same topic. In fact, we probably are not in the same field.
- Jovan works on scrutibility in personalized time management systems
- Dipto Barman works on fake news and exposure personalization
- Hassan Zaal works on care-giving human-robot interaction for elderly patients
- Awais Akbar works on route pathing and planning as per user preferences
How did we resolve this? Well, we did a one-minute rapid thesis presentation of our work while noting down the keywords that were the most prominent towards a subject, area, or topic. We then analysed the relationship between our topics and realized that one interesting common thread between all our work was the relevance and importance of context. In each of our work, representing and considering contextual information changes the way the agent would make decisions and generally impacts agent behaviour and its appropriateness in the context of the interaction.
Topics of Interest
Once it was defined that we were setting up a workshop that deals with the representation of context in intelligent systems which was broadly within the domina of user modeling, adaptive systems, and personalization, the next question was how to choose the relevant topics of interest. Thankfully, we had help from conferences and workshops surrounding the study of contextual information in recommender systems and other avenues of ubiquitous computing. We decided to increase our focus on scrutibility as a key focus of context and relevance for our topics of interest for the workshop.
Overall, the lens from which context was contextualized was to look at user-centric recommender-like systems and how context might play a role if correctly captured and used.
Papers and Actual Workshop
Featuring four accepted papers out of five submitted, and being one of teh first workshops to showcase the divergent viewpoints in scrutibility within the confines of an atypical theoretical domain of context relevance, this workshop was an avenue for researchers of differnt verticles of user-based design and computing research to understand context.
The workshop featured papers from ubiquitous computing for health, route planning, and hierarchical recommender systems, even in its smaller number of submissions. The format of the workshop included a keynote by Co-chair Prof. Judy Kay, the paper presentations, a panel disucssion, and a joint workshop brainstorming activity
Prof. Masthoff started the program with her opening talk with a presentation which explored and analyzed the role of context in understanding systems and scrutinizing them, which became a topic of conversation for the rest of the workshop. The panel, comprising of speakers Prof. Judith Masthoff, Dr. Barakat A. Yilma, Dr. David Massimo, and co-chair Prof. Owen Conlan (which I was not able to capture cuz it was not hybrid and I did not travel to Cyprus).
Final Thoughts
Organizing this workshop was fun. While I think having two former general chairs of the conference our workshop was co-located with definitely helps, I view it as a well-deserved success. The workshop fostered spirited academic debate, and is shaping up to be a venue which provides value to the user modeling as well as the analysis of computing community at large.