When and Where:
- September 9, 2016 from 1:30-2:30 PM (1 hour)
- 777 West Main Street, Boise, ID 83702
- (CCP) City Center Plaza Room 221
- All are invited, no registration required
- Contact email@example.com for additional info
Dr. Martijn Willemsen is an expert on human decision making in interactive systems. He works in the Human-Technology Interaction group of Eindhoven University of Technology (The Netherlands). His primary interests lie in the understanding of cognitive processes of decision making by means of process tracing and in the application of decision making theory in interactive systems such as recommender systems. He is also an expert on user-centric evaluation of adaptive systems.
Topic: What Recommender Systems Can Learn from Decision Psychology about Preference Elicitation & Behavioral Change
Recommender systems typically use collaborative filtering: information from your preferences (i.e. your ratings) is combined with that of other users to predict what other items you might also like. Much research in the field has focused on building algorithms that provide the most accurate recommendations. However, these models make strong assumptions about how preferences come about, how stable they are, and how they can be measured. I have studied how the ways recommender systems learn your preferences (preference elicitation) can be better understood and improved based on psychological insights. I will discuss how our memory influences our ratings, why ratings (as an absolute measure of preference) have issues, and recent work on other types of preference elicitation that use relative measures such as choice rather than rating.
Moreover, I will discuss new ideas on recommending for behavioral change: when people want to improve their behavior (become more sustainable, live a healthier life) we need different algorithms that do not predict what we choose or do now but what we should choose to improve on our behavior (Ekstrand and Willemsen, 2016). I will present an example of such an alternative approach to recommendations using a Rasch scale, which orders behaviors in terms of their difficulty or costs and models a user’s ability to perform these behaviors, with user studies of applying this scale in two different systems that provide personalized suggestions for energy-savings and for reducing hypertension.