{"slug": "auditing-the-risk-claims-of-distributional-reinforcement-learning", "title": "Auditing the Risk Claims of Distributional Reinforcement Learning", "summary": "A new audit of distributional reinforcement learning agents (QR-DQN, C51, IQN) on MinAtar and Atari games finds that 40-95% of the strongest claimed risk trade-offs are refuted at 95% confidence, with the learned \"risk\" reflecting a training artifact rather than environment stochasticity. The artifact is structural, appears unchanged at full-Atari scale, and persists even with risk training or ensembling, indicating that the agents' risk claims are uninformative.", "body_md": "# Computer Science > Artificial Intelligence\n\n[Submitted on 13 Jul 2026]\n\n# Title:Auditing the Risk Claims of Distributional Reinforcement Learning\n\n[View PDF](/pdf/2607.11607)\n\n[HTML (experimental)](https://arxiv.org/html/2607.11607v1)\n\nAbstract:Distributional reinforcement learning agents learn full return distributions that are increasingly read at face value: for interpretability, risk-sensitive control, and safety monitoring. We ask a question theory anticipates but that has not been measured directly: are the risk claims of a trained distributional agent true? Our audit combines a decision-relevant screening metric (the excess Wasserstein gap between the top two actions, which equals the mass by which first-order stochastic dominance is violated), ground truth from snapshot-restart Monte Carlo, and a statistical harness (permutation nulls, bootstrap refutation, FDR control) without which the audit itself manufactures false conclusions. Across QR-DQN, C51, and IQN on MinAtar (33 runs), 40-95% of the strongest claimed risk trade-offs are refuted at 95% confidence, the placement of the strongest claims is statistically indistinguishable from truth-blind, and essentially no claim is confirmable: for these agents, the learned \"risk\" reflects a training artifact rather than environment stochasticity. The artifact is structural (fully formed early in training, uncorrelated with final score, idiosyncratic to each seed) and appears unchanged at full-Atari scale, with every top Breakout claim of a pretrained near-state-of-the-art QR-DQN refuted. Positive controls of known magnitude confirm 96-100% of real claims (correlation 0.89-0.92): the reading measures the agents, not the audit. Acting on the heads' CVaR advice at their most-flagged states ranges from beneficial to significantly worse than chance. Neither training for risk nor ensembling removes the artifact, and recalibration passes the audit only by nullifying the claims: the head is uninformative, not merely miscalibrated. We release the toolkit and document two silent pitfalls that produced convincing but wrong audits of our own.\n\n### Current browse context:\n\ncs.AI\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/auditing-the-risk-claims-of-distributional-reinforcement-learning", "canonical_source": "https://arxiv.org/abs/2607.11607", "published_at": "2026-07-14 12:59:11+00:00", "updated_at": "2026-07-14 13:18:22.444651+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "ai-safety", "ai-ethics"], "entities": ["QR-DQN", "C51", "IQN", "MinAtar", "Atari", "Breakout"], "alternates": {"html": "https://wpnews.pro/news/auditing-the-risk-claims-of-distributional-reinforcement-learning", "markdown": "https://wpnews.pro/news/auditing-the-risk-claims-of-distributional-reinforcement-learning.md", "text": "https://wpnews.pro/news/auditing-the-risk-claims-of-distributional-reinforcement-learning.txt", "jsonld": "https://wpnews.pro/news/auditing-the-risk-claims-of-distributional-reinforcement-learning.jsonld"}}