Abstract: 

Recommendation systems occupy an expanding role in everyday decision making; they have the potential to influence product consumption, individuals' perceptions of the world, and life-altering medical and legal decisions. The data used to train and evaluate these systems is algorithmically confounded: users are already exposed to algorithmic recommendations, creating a feedback loop between human choices and the recommendation system. Using simulations, we will demonstrate how using data confounded in this way can impact both individuals and the platform as a whole.

Bio: 

Allison Chaney is interested in understanding and characterizing the impact of machine learning models and algorithms on people, both individually and societally, and in developing new machine learning methods to improve not only companies' objectives, but also individual well-being and societal welfare.