TY - JOUR
T1 - Learning latent structure
T2 - Carving nature at its joints
AU - Gershman, Samuel J.
AU - Niv, Yael
N1 - Funding Information:
We are grateful to Michael Todd for many helpful discussions. SJG is funded by a Quantitative Computational Neuroscience training grant from the National Institute of Mental Health.
PY - 2010/4
Y1 - 2010/4
N2 - 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.
AB - 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.
UR - http://www.scopus.com/inward/record.url?scp=77952541839&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77952541839&partnerID=8YFLogxK
U2 - 10.1016/j.conb.2010.02.008
DO - 10.1016/j.conb.2010.02.008
M3 - Review article
C2 - 20227271
AN - SCOPUS:77952541839
SN - 0959-4388
VL - 20
SP - 251
EP - 256
JO - Current Opinion in Neurobiology
JF - Current Opinion in Neurobiology
IS - 2
ER -