We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Machine Learning

Title: Optimistic Risk Perception in the Temporal Difference error Explains the Relation between Risk-taking, Gambling, Sensation-seeking and Low Fear

Abstract: Understanding the affective, cognitive and behavioural processes involved in risk taking is essential for treatment and for setting environmental conditions to limit damage. Using Temporal Difference Reinforcement Learning (TDRL) we computationally investigated the effect of optimism in risk perception in a variety of goal-oriented tasks. Optimism in risk perception was studied by varying the calculation of the Temporal Difference error, i.e., delta, in three ways: realistic (stochastically correct), optimistic (assuming action control), and overly optimistic (assuming outcome control). We show that for the gambling task individuals with 'healthy' perception of control, i.e., action optimism, do not develop gambling behaviour while individuals with 'unhealthy' perception of control, i.e., outcome optimism, do. We show that high intensity of sensations and low levels of fear co-occur due to optimistic risk perception. We found that overly optimistic risk perception (outcome optimism) results in risk taking and in persistent gambling behaviour in addition to high intensity of sensations. We discuss how our results replicate risk-taking related phenomena.
Subjects: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
Cite as: arXiv:1404.2078 [cs.LG]
  (or arXiv:1404.2078v2 [cs.LG] for this version)

Submission history

From: Joost Broekens [view email]
[v1] Tue, 8 Apr 2014 10:26:27 GMT (1098kb)
[v2] Tue, 3 Feb 2015 14:09:51 GMT (1251kb)

Link back to: arXiv, form interface, contact.