{
  "abstract": "Modern multi-agent systems are increasingly deployed at scale over large populations of agents in settings such as ad-auctions, traffic routing, and recommendation systems. The dominant approach in such settings is to optimize each agent's policy independently, treating the other agents as part of a fixed single-agent environment rather than modeling the population dynamics. In many large-population systems, the dynamics depend on an aggregate summary of the population rather than the identity of any individual. Mean-field RL exploits such structure, providing a principled framework that models each agent's environment as an explicit function of the population distribution. However, in large state-action spaces or high-dimensional control problems, modeling the population distribution is itself intractable. How can we design a scalable framework for high-dimensional control problems with large populations? This work explores this question from the perspective of representation learning. We introduce a mean-field RL framework in which the rewards and transition dynamics depend on the population only through an unknown low-dimensional aggregate statistic. We then study this framework in the offline setting and design a provable approach that learns a near-optimal policy by learning a low-dimensional representation. Motivated by real-life supply-chain optimization problems, we design a one-step routing game to test the hypothesis that learning a low-dimensional population representation improves reward prediction and Nash gap estimation relative to baselines that don't exploit this structure. We show that under a fixed neural-network parameter count and optimization budget, learning a low-dimensional population representation improves reward prediction and the equilibrium quality of the resulting policies.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.02928",
  "author_total": 6,
  "authors": [
    {
      "name": "Aditya Makkar"
    },
    {
      "name": "Benjamin Unger"
    },
    {
      "name": "Jeongyeol Kwon"
    },
    {
      "name": "Mathieu Laurière"
    },
    {
      "name": "Eugene Vinitsky"
    },
    {
      "name": "Yonathan Efroni"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/bb3bx8dhwzebt.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.02928",
  "fulltext_url": null,
  "id": "bb3bx8dhwzebt",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.02928",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "cs.MA"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1702|Reinforcement Learning in Robotics"
    ]
  ],
  "openalex": "W7208776003",
  "pdf_url": "https://arxiv.org/pdf/2609.02928",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-01",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/bb3bx8dhwzebt.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:18:42Z",
      "native_id": "2609.02928",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.02928"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Towards Scaling Reinforcement Learning to Massive Populations: Learning Mean-Field Representations",
  "title_text": "Towards Scaling Reinforcement Learning to Massive Populations: Learning Mean-Field Representations",
  "type": "preprint",
  "updated_at": "2026-10-30T12:18:42Z",
  "url": "https://research.draconitic.com/paper/bb3bx8dhwzebt",
  "venue": "arXiv",
  "versions": []
}