Deviation Ratings: A General, Clone-Invariant Rating Method

• URL: http://arxiv.org/abs/2502.11645v1
• Date: Mon, 17 Feb 2025 10:39:04 GMT
• Title: Deviation Ratings: A General, Clone-Invariant Rating Method
• Authors: Luke Marris, Siqi Liu, Ian Gemp, Georgios Piliouras, Marc Lanctot,
• Abstract summary: This work introduces the first N-player general-sum clone invariant rating, called deviation ratings, based on coarse correlated equilibria. The rating is explored on several domains including LLMs evaluation.
• Score: 39.480611712794094
• License:
• Abstract: Many real-world multi-agent or multi-task evaluation scenarios can be naturally modelled as normal-form games due to inherent strategic (adversarial, cooperative, and mixed motive) interactions. These strategic interactions may be agentic (e.g. players trying to win), fundamental (e.g. cost vs quality), or complementary (e.g. niche finding and specialization). In such a formulation, it is the strategies (actions, policies, agents, models, tasks, prompts, etc.) that are rated. However, the rating problem is complicated by redundancy and complexity of N-player strategic interactions. Repeated or similar strategies can distort ratings for those that counter or complement them. Previous work proposed ``clone invariant'' ratings to handle such redundancies, but this was limited to two-player zero-sum (i.e. strictly competitive) interactions. This work introduces the first N-player general-sum clone invariant rating, called deviation ratings, based on coarse correlated equilibria. The rating is explored on several domains including LLMs evaluation.

Related papers

• Toward Optimal LLM Alignments Using Two-Player Games [86.39338084862324]
  In this paper, we investigate alignment through the lens of two-agent games, involving iterative interactions between an adversarial and a defensive agent. We theoretically demonstrate that this iterative reinforcement learning optimization converges to a Nash Equilibrium for the game induced by the agents. Experimental results in safety scenarios demonstrate that learning in such a competitive environment not only fully trains agents but also leads to policies with enhanced generalization capabilities for both adversarial and defensive agents.
  arXiv  Detail & Related papers  (2024-06-16T15:24:50Z)
• Paths to Equilibrium in Games [6.812247730094933]
  We study sequences of strategies satisfying a pairwise constraint inspired by policy updating in reinforcement learning. Our analysis reveals a counterintuitive insight that reward deteriorating strategic updates are key to driving play to equilibrium along a satisficing path.
  arXiv  Detail & Related papers  (2024-03-26T19:58:39Z)
• Strategic Apple Tasting [35.25249063553063]
  Algorithmic decision-making in high-stakes domains often involves assigning decisions to agents with incentives to strategically modify their input to the algorithm. We formalize this setting as an online learning problem with apple-tasting feedback. Our goal is to achieve sublinear strategic regret, which compares the performance of the principal to that of the best fixed policy in hindsight.
  arXiv  Detail & Related papers  (2023-06-09T20:46:31Z)
• On the Complexity of Multi-Agent Decision Making: From Learning in Games to Partial Monitoring [105.13668993076801]
  A central problem in the theory of multi-agent reinforcement learning (MARL) is to understand what structural conditions and algorithmic principles lead to sample-efficient learning guarantees. We study this question in a general framework for interactive decision making with multiple agents. We show that characterizing the statistical complexity for multi-agent decision making is equivalent to characterizing the statistical complexity of single-agent decision making.
  arXiv  Detail & Related papers  (2023-05-01T06:46:22Z)
• Opponent Modeling in Multiplayer Imperfect-Information Games [1.024113475677323]
  We present an approach for opponent modeling in multiplayer imperfect-information games. We run experiments against a variety of real opponents and exact Nash equilibrium strategies in three-player Kuhn poker. Our algorithm significantly outperforms all of the agents, including the exact Nash equilibrium strategies.
  arXiv  Detail & Related papers  (2022-12-12T16:48:53Z)
• Finding mixed-strategy equilibria of continuous-action games without gradients using randomized policy networks [83.28949556413717]
  We study the problem of computing an approximate Nash equilibrium of continuous-action game without access to gradients. We model players' strategies using artificial neural networks. This paper is the first to solve general continuous-action games with unrestricted mixed strategies and without any gradient information.
  arXiv  Detail & Related papers  (2022-11-29T05:16:41Z)
• Game Theoretic Rating in N-player general-sum games with Equilibria [26.166859475522106]
  We propose novel algorithms suitable for N-player, general-sum rating of strategies in normal-form games according to the payoff rating system. This enables well-established solution concepts, such as equilibria, to be leveraged to efficiently rate strategies in games with complex strategic interactions.
  arXiv  Detail & Related papers  (2022-10-05T12:33:03Z)
• Provably Efficient Fictitious Play Policy Optimization for Zero-Sum Markov Games with Structured Transitions [145.54544979467872]
  We propose and analyze new fictitious play policy optimization algorithms for zero-sum Markov games with structured but unknown transitions. We prove tight $widetildemathcalO(sqrtK)$ regret bounds after $K$ episodes in a two-agent competitive game scenario. Our algorithms feature a combination of Upper Confidence Bound (UCB)-type optimism and fictitious play under the scope of simultaneous policy optimization.
  arXiv  Detail & Related papers  (2022-07-25T18:29:16Z)
• Provably Efficient Offline Multi-agent Reinforcement Learning via Strategy-wise Bonus [48.34563955829649]
  We propose a strategy-wise concentration principle which builds a confidence interval for the joint strategy. For two-player zero-sum Markov games, by exploiting the convexity of the strategy-wise bonus, we propose an efficient algorithm. All of our algorithms can naturally take a pre-specified strategy class $Pi$ as input and output a strategy close to the best strategy in $Pi$.
  arXiv  Detail & Related papers  (2022-06-01T00:18:15Z)
• On the Impossibility of Convergence of Mixed Strategies with No Regret Learning [10.515544361834241]
  We study convergence properties of the mixed strategies that result from a general class of optimal no regret learning strategies. We consider the class of strategies whose information set at each step is the empirical average of the opponent's realized play.
  arXiv  Detail & Related papers  (2020-12-03T18:02:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.