Seminar: Similarity and Consistency in Algorithm-Guided Exploration

© Bastian Dincher, University of Bremen

Sebastian Fehrler, Universität Bremen

Similarity and Consistency in Algorithm-Guided Exploration

Algorithm-based decision support systems are crucial for exploration tasks like product searches and portfolio choices. In an online study, we examine if participants' acceptance of a reinforcement learning algorithm's advice hinges on their exploration preferences aligning with the algorithm's. We adjust the algorithm's emphasis on exploration versus exploitation and model participants' decision-making processes accordingly. Results show participants are more inclined to follow an algorithm favoring exploitation, possibly perceiving its consistency as expertise. However, similarity between human choices and algorithmic recommendations doesn't increase willingness to follow advice. Thus, consistency in an algorithm's recommendations over time is pivotal in encouraging people to heed algorithmic advice in exploration tasks.

Sebastian Fehrler, Professor of Economics at the University of Bremen's SOCIUM and contributor to the Humans on Mars Initiative, studies in human-machine interactions, alongside research in communication, group decision-making, and evaluating social-policy interventions. His approach incorporates game theory and experimental methods (lab and field). Previously, he earned his PhD at the University of Zurich, completed a visiting postdoc year at NYU, and served as an Assistant Professor of Behavioral Economics at the University of Konstanz.

Join at the Max Planck Institute for Human Development or online.
Webex Access
Meeting number: 2744 343 5034
Password: E4eK5VcpW8c