Multi-agent Communication meets Natural Language: Synergies between Functional and Structural Language Learning

Lazaridou, Angeliki; Potapenko, Anna; Tieleman, Olivier

doi:10.48550/arxiv.2005.07064

Cited by 2 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MARL has been used to study a variety of social situations, including social dilemmas (Hughes et al, 2018;Lerer and Peysakhovich, 2017), norm and convention formation (Bakker et al, 2021;Köster et al, 2020;Lerer and Peysakhovich, 2019), communication (Foerster et al, 2016;Lazaridou et al, 2020), and many more. A recent article explored partner choice in an abstract iterated social dilemma environment (Anastassacos et al, 2020), validating that partner choice can indeed promote cooperation in learning agents.…”

Section: Reinforcement Learning As a Model Of Human Behaviormentioning

confidence: 99%

Statistical discrimination in learning agents

Duéñez-Guzmán¹,

McKee²,

Mao³

et al. 2021

Preprint

View full text Add to dashboard Cite

Independent researcherUndesired bias afflicts both human and algorithmic decision making, and may be especially prevalent when information processing trade-offs incentivize the use of heuristics. One primary example is statistical discrimination-selecting social partners based not on their underlying attributes, but on readily perceptible characteristics that covary with their suitability for the task at hand. We present a theoretical model to examine how information processing influences statistical discrimination and test its predictions using multi-agent reinforcement learning with various agent architectures in a partner choice-based social dilemma. As predicted, statistical discrimination emerges in agent policies as a function of both the bias in the training population and of agent architecture. All agents showed substantial statistical discrimination, defaulting to using the readily available correlates instead of the outcome relevant features. We show that less discrimination emerges with agents that use recurrent neural networks, and when their training environment has less bias. However, all agent algorithms we tried still exhibited substantial bias after learning in biased training populations.

show abstract

Section: Reinforcement Learning As a Model Of Human Behaviormentioning

confidence: 99%

Statistical discrimination in learning agents

Duéñez-Guzmán¹,

McKee²,

Mao³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The introduction of complex environments to study agent strategies and communication was a key advance in multiagent reinforcement learning research [18,19,22,26]. Specifically, there has been a significant body of work on temporally-extended games to understand interesting social aspects of interactions [12,26], agents [34], social dilemmas [19], cooperative games [6], and competitive games [2].…”

Section: Introductionmentioning

confidence: 99%

Hidden Agenda: a Social Deduction Game with Diverse Learned Equilibria

Kopparapu¹,

Duéñez-Guzmán²,

Matyas³

et al. 2022

Preprint

View full text Add to dashboard Cite

A key challenge in the study of multiagent cooperation is the need for individual agents not only to cooperate effectively, but to decide with whom to cooperate. This is particularly critical in situations when other agents have hidden, possibly misaligned motivations and goals. Social deduction games offer an avenue to study how individuals might learn to synthesize potentially unreliable information about others, and elucidate their true motivations. In this work, we present Hidden Agenda, a two-team social deduction game that provides a 2D environment for studying learning agents in scenarios of unknown team alignment. The environment admits a rich set of strategies for both teams. Reinforcement learning agents trained in Hidden Agenda show that agents can learn a variety of behaviors, including partnering and voting without need for communication in natural language.

show abstract

Multi-agent Communication meets Natural Language: Synergies between Functional and Structural Language Learning

Cited by 2 publications

References 13 publications

Statistical discrimination in learning agents

Statistical discrimination in learning agents

Hidden Agenda: a Social Deduction Game with Diverse Learned Equilibria

Contact Info

Product

Resources

About