TY - JOUR
T1 - Optimal Behavioral Hierarchy
AU - Solway, Alec
AU - Diuk, Carlos
AU - Córdova, Natalia
AU - Yee, Debbie
AU - Barto, Andrew G.
AU - Niv, Yael
AU - Botvinick, Matthew M.
N1 - Funding Information:
James S. McDonnell Foundation (http://www.jsmf.org/). National Science Foundation (#1207833) (nsf.gov). John Templeton Foundation (http://www. templeton.org/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Publisher Copyright:
© 2014 Solway et al.
PY - 2014/8/14
Y1 - 2014/8/14
N2 - Human behavior has long been recognized to display hierarchical structure: actions fit together into subtasks, which cohere into extended goal-directed activities. Arranging actions hierarchically has well established benefits, allowing behaviors to be represented efficiently by the brain, and allowing solutions to new tasks to be discovered easily. However, these payoffs depend on the particular way in which actions are organized into a hierarchy, the specific way in which tasks are carved up into subtasks. We provide a mathematical account for what makes some hierarchies better than others, an account that allows an optimal hierarchy to be identified for any set of tasks. We then present results from four behavioral experiments, suggesting that human learners spontaneously discover optimal action hierarchies.
AB - Human behavior has long been recognized to display hierarchical structure: actions fit together into subtasks, which cohere into extended goal-directed activities. Arranging actions hierarchically has well established benefits, allowing behaviors to be represented efficiently by the brain, and allowing solutions to new tasks to be discovered easily. However, these payoffs depend on the particular way in which actions are organized into a hierarchy, the specific way in which tasks are carved up into subtasks. We provide a mathematical account for what makes some hierarchies better than others, an account that allows an optimal hierarchy to be identified for any set of tasks. We then present results from four behavioral experiments, suggesting that human learners spontaneously discover optimal action hierarchies.
UR - http://www.scopus.com/inward/record.url?scp=84923228672&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84923228672&partnerID=8YFLogxK
U2 - 10.1371/journal.pcbi.1003779
DO - 10.1371/journal.pcbi.1003779
M3 - Article
C2 - 25122479
AN - SCOPUS:84923228672
SN - 1553-734X
VL - 10
JO - PLoS computational biology
JF - PLoS computational biology
IS - 8
M1 - e1003779
ER -