{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/multi-agent-reinforcement-learning-in","title":"Multi-agent Reinforcement Learning in Sequential Social Dilemmas","arxiv_id":"1702.03037","date":"2017-02-10","proceeding":null,"authors":["Joel Z. Leibo","Vinicius Zambaldi","Marc Lanctot","Janusz Marecki","Thore Graepel"],"abstract":"Matrix games like Prisoner's Dilemma have guided research on social dilemmas\nfor decades. However, they necessarily treat the choice to cooperate or defect\nas an atomic action. In real-world social dilemmas these choices are temporally\nextended. Cooperativeness is a property that applies to policies, not\nelementary actions. We introduce sequential social dilemmas that share the\nmixed incentive structure of matrix game social dilemmas but also require\nagents to learn policies that implement their strategic intentions. We analyze\nthe dynamics of policies learned by multiple self-interested independent\nlearning agents, each using its own deep Q-network, on two Markov games we\nintroduce here: 1. a fruit Gathering game and 2. a Wolfpack hunting game. We\ncharacterize how learned behavior in each domain changes as a function of\nenvironmental factors including resource abundance. Our experiments show how\nconflict can emerge from competition over shared resources and shed light on\nhow the sequential nature of real world social dilemmas affects cooperation.","url_abs":"http://arxiv.org/abs/1702.03037v1","url_pdf":"http://arxiv.org/pdf/1702.03037v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"multi-agent-reinforcement-learning-in","repo_url":"https://github.com/eugenevinitsky/sequential_social_dilemma_games","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"multi-agent-reinforcement-learning-in","repo_url":"https://github.com/quantaji/AMD-SSD","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"multi-agent-reinforcement-learning-in","repo_url":"https://github.com/social-dilemma/multiagent","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"multi-agent-reinforcement-learning-in","repo_url":"https://github.com/vermashresth/sequential_social_dilemma_games","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"multi-agent-reinforcement-learning","task_name":"Multi-agent Reinforcement Learning"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1702.03037","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}