Mining Relevance from Human Neuro-physiology

Abstract

User behavior analysis drive today's internet applications. The pages we visit, the amount of time we view each page, the links we click on, the searches we make, and the things that we interact with, allow collecting data that can be used to improve web search, recommendation systems, along with targeting online advertisement for users who exhibit a greater level of interest and intent for the information being returned. While user behavior has been successfully utilized for improving services, it is limited to the measurements that the current user interfaces allow us to record. In this talk, I will present some of our recent work on going beyond the behavioral measures to utilize human physiology and brain-activity to detect and mine relevance and interest towards digital content.

I will discuss two cases in which peripheral physiology and EEG are used for recording user signals, and features inferred from these signals are utilized to classify relevance and user's interest towards digital content. I will also discuss the research challenges that are ahead in utilizing this technology outside research labs.

Bio: Tuukka Ruotsalo's research areas span human-computer interaction, information retrieval, and applied machine learning to elicit and mine user preferences and intentions to drive novel information discovery applications. Tuukka is currently an academy of Finland research fellow at Helsinki Institute for Information Technology in Finland, and a visiting scholar at Harvard SEAS, Computer Science. He has previously been affiliated with University of Helsinki, Finland, the Free University of Amsterdam in the Netherlands, and the University of California in Berkeley. Tuukka is also a co-founder of Etsimo, a start-up that builds information discovery solutions.