Learning from other minds: an optimistic critique of reinforcement learning models of social learning. (April 2021)

• Record Type: Journal Article
• Authors: Vélez, Natalia; Gweon, Hyowon
• Is Part Of: Current opinion in behavioral sciences. Volume 38(2021)
• Page Start: 110
• Page End: 115
• Publication Date: 2021-04
• Subjects: Psychology -- Periodicals; 150.5
• DOI: 10.1016/j.cobeha.2021.01.006
• Languages: English
• ISSNs: 2352-1546

Abstract

Highlights: RL models have been applied to study the neural underpinnings of social learning. Past work has largely found neural correlates of observable, reward-predictive cues. But, in learning from others, humans make inferences that go beyond observable data. Such inferences rely on a model of how others' mental states give rise to actions. Bridging RL and Bayesian approaches can deepen our understanding of social learning. Abstract : Reinforcement learning models have been productively applied to identify neural correlates of the value of social information. However, by operationalizing social information as a lean, reward-predictive cue, this literature underestimates the richness of human social learning: Humans readily go beyond action-outcome mappings and can draw flexible inferences from a single observation. We argue that computational models of social learning need minds, that is, a generative model of how others' unobservable mental states cause their observable actions. Recent advances in inferential social learning suggest that even young children learn from others by using an intuitive, generative model of other minds. Bridging developmental, Bayesian, and reinforcement learning perspectives can enrich our understanding of the neural bases of distinctively human social learning.