Citation
R. Neumann, A. N. Rafferty, T. L. Griffiths. "A Bounded Rationality Account of Wishful Thinking". Annual Conference of the Cognitive Science Society, 2014.

Abstract
People tend towards wishful thinking, in which they overestimate the probability of favorable outcomes and underestimate the probability of unfavorable outcomes. Many explanations for this phenomenon focus on its irrationality. We explore whether wishful thinking could actually help people make better decisions given that they have limited cognitive resources. We consider a situation in which multiple decisions must be made over a period of time, where the consequences of these decisions are not fully determined. We model this situation as a Markov decision process, and incorporate limited cognitive resources by varying the amount of time in the future that the agent considers the consequences of its decisions. Through simulations, we show that with limited cognitive resources, this model can exhibit better performance by incorporating a bias towards wishful thinking. This advantage occurs across a range of decision-making environments, suggesting that the same effect could be applicable to many real life scenarios.

Electronic downloads

• http://cocosci.berkeley.edu/anna/papers/WishfulThinking.pdf

• HTML

  R. Neumann, A. N. Rafferty, T. L. Griffiths. <a
  href="http://robotics.eecs.berkeley.edu/pubs/14.html"
  >A Bounded Rationality Account of Wishful
  Thinking</a>, Annual Conference of the Cognitive
  Science Society, 2014.

• Plain text

  R. Neumann, A. N. Rafferty, T. L. Griffiths. "A Bounded
  Rationality Account of Wishful Thinking". Annual
  Conference of the Cognitive Science Society, 2014.

• BibTeX

  @inproceedings{NeumannRaffertyGriffiths14_BoundedRationalityAccountOfWishfulThinking,
      author = {R. Neumann and A. N. Rafferty and T. L. Griffiths},
      title = {A Bounded Rationality Account of Wishful Thinking},
      booktitle = {Annual Conference of the Cognitive Science Society},
      year = {2014},
      abstract = {People tend towards wishful thinking, in which
                they overestimate the probability of favorable
                outcomes and underestimate the probability of
                unfavorable outcomes. Many explanations for this
                phenomenon focus on its irrationality. We explore
                whether wishful thinking could actually help
                people make better decisions given that they have
                limited cognitive resources. We consider a
                situation in which multiple decisions must be made
                over a period of time, where the consequences of
                these decisions are not fully determined. We model
                this situation as a Markov decision process, and
                incorporate limited cognitive resources by varying
                the amount of time in the future that the agent
                considers the consequences of its decisions.
                Through simulations, we show that with limited
                cognitive resources, this model can exhibit better
                performance by incorporating a bias towards
                wishful thinking. This advantage occurs across a
                range of decision-making environments, suggesting
                that the same effect could be applicable to many
                real life scenarios.},
      URL = {http://robotics.eecs.berkeley.edu/pubs/14.html}
  }
  

Posted by Ehsan Elhamifar on 30 May 2014.
Groups: ehumans
For additional information, see the Publications FAQ or contact webmaster at robotics eecs berkeley edu.