English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The deployment of machine learning systems in the market economy hastriggered academic and institutional fears over potential tacit collusionbetween fully automated agents. Multiple recent economics studies haveempirically shown the emergence of collusive strategies from agents guided bymachine learning algorithms. In this work, we prove that multi-agent Q-learnersplaying the iterated prisoner's dilemma can learn to collude. The complexity ofthe cooperative multi-agent setting yields multiple fixed-point policies for$Q$-learning: the main technical contribution of this work is to characterizethe convergence towards a specific cooperative policy. More precisely, in theiterated prisoner's dilemma, we show that with optimistic Q-values, anyself-play Q-learner can provably learn a cooperative policy called Pavlov, alsoreferred to as win-stay, lose-switch policy, which strongly differs from thevanilla Pareto-dominated always defect policy.

Q-learners Can Provably Collude in the Iterated Prisoner's Dilemma