Surprise, Curiosity and Reward: An Evolutionary Perspective

This lecture explores the evolutionary perspective on surprise, curiosity, and reward, focusing on the role of primary and secondary reward signals in influencing brain activity and decision-making. The speaker discusses the Reward-Prediction-Error Hypothesis and the impact of meta-optimization experiments on reinforcement learning. The presentation delves into the challenges of designing effective reward signals and the importance of intrinsic motivation in guiding behavior. Additionally, the lecture examines the concept of curiosity and its relationship to exploration behavior, drawing on insights from foundational research in the field. The session concludes with a reflection on the implications of surprise and value in shaping adaptive decision-making processes.