skip to main content

Behavioral Social Neuroscience Seminar

Wednesday, March 5, 2014
4:00pm to 5:00pm
Add to Cal
Beckman Behavioral Biology B180
Computational Heuristics in Planning
Quentin Huys, University Hospital of Psychiatry Zurich,

Humans routinely formulate plans in domains whose complexity challenges even the most powerful computers. To do so, they appear to avail themselves of a plethora of strategies and heuristics that efficiently simplify, approximate and hierarchically decompose hard tasks into simpler subtasks. Theoretical and cognitive research has suggested and provided evidence for particular strategies; however, little is known about their establishment, interaction and efficiency. Here, we use model-based behavioural analysis to provide a forensic examination of the detailed performance of human subjects in a moderately deep planning task. We find that subjects exploit the structure of the domain to establish subgoals in a way that achieves a near maximal reduction in the cost of computing the value of choices, but then combine partial searches with greedy local steps to solve subtasks, and maladaptively prune the decision trees of subtasks in a reflexive manner upon encountering salient losses. Subjects come idiosyncratically to favour particular sequences of actions to achieve subgoals, creating novel complex actions or 'options'.

For more information, please contact Barbara Estrada by phone at Ext. 4083 or by email at [email protected].