Learning latent structure: Carving nature at its joints

Samuel J. Gershman, Yael Niv

Research output: Contribution to journalReview article

106 Scopus citations

Abstract

Reinforcement learning (RL) algorithms provide powerful explanations for simple learning and decision-making behaviors and the functions of their underlying neural substrates. Unfortunately, in real-world situations that involve many stimuli and actions, these algorithms learn pitifully slowly, exposing their inferiority in comparison to animal and human learning. Here we suggest that one reason for this discrepancy is that humans and animals take advantage of structure that is inherent in real-world tasks to simplify the learning problem. We survey an emerging literature on 'structure learning'. -. using experience to infer the structure of a task. -. and how this can be of service to RL, with an emphasis on structure in perception and action.

Original languageEnglish (US)
Pages (from-to)251-256
Number of pages6
JournalCurrent Opinion in Neurobiology
Volume20
Issue number2
DOIs
StatePublished - Apr 2010

All Science Journal Classification (ASJC) codes

  • Neuroscience(all)

Fingerprint Dive into the research topics of 'Learning latent structure: Carving nature at its joints'. Together they form a unique fingerprint.

  • Cite this