Naturalistic learning and decision making reading group

This is a summer reading group in the Gershman lab.

Real-world decisions feature numerous complexities that are seldom captured by controlled laboratory settings. Environmental states are high-dimensional and contain task-irrelevant information, the agents themselves must nontrivially synthesize a set of considered actions, while contextual information may encourage the elicitation of past experiences and enable their partial transfer to the current task. In our reading group, we will survey theoretical and empirical works that characterize human learning and decision making in such naturalistic settings, and discuss fruitful ways to reach deeper understanding. The presented papers may come from various traditions including cognitive science, machine learning, economics, and field studies.