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

Download:

Current browse context:

q-bio.PE

Change to browse by:

References & Citations

Bookmark

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

Quantitative Biology > Populations and Evolution

Title: Evolutionary (in)stability of selfish learning in repeated games

Abstract: Evolutionary game theory offers a general framework to describe how humans revise their behavior in strategic environments. To model this adaptation process, most previous work assumes that individuals aim to increase their own short-run payoff. This kind of "selfish learning," however, entails the risk of getting trapped in equilibria that are detrimental to everyone. If evolution operates on the level of long-run payoffs, it might thus select for a different learning rule altogether. Motivated by experimental work, we therefore study an alternative rule called "fairness-mediated team learning" (FMTL). An FMTL learner aims to maximize the group payoff while minimizing payoff differences between group members. When adopted by everyone, FMTL is superior to selfish learning, both individually and socially, across many different social dilemmas. Remarkably, however, we show that FMTL exhibits a similar performance even against selfish learners. Based on these observations, we explore the dynamics that arise when the two learning rules themselves are subject to evolution. If individuals are sufficiently patient to consider the long-run consequences of their learning rules, selfish learning is routinely invaded. These results further corroborate previous theoretical attempts to explain why humans take into account their impact on others when making strategic decisions.
Comments: 56 pages
Subjects: Populations and Evolution (q-bio.PE)
Cite as: arXiv:2105.06199 [q-bio.PE]
  (or arXiv:2105.06199v1 [q-bio.PE] for this version)

Submission history

From: Alex McAvoy [view email]
[v1] Thu, 13 May 2021 11:37:09 GMT (9710kb,D)
[v2] Mon, 30 Aug 2021 15:22:19 GMT (9773kb,D)
[v3] Wed, 31 Aug 2022 23:24:53 GMT (10996kb,AD)

Link back to: arXiv, form interface, contact.