Learning Strategic Language Agents in the Werewolf Game with Iterative Latent Space Policy Optimization

• URL: http://arxiv.org/abs/2502.04686v3
• Date: Wed, 18 Jun 2025 03:24:18 GMT
• Title: Learning Strategic Language Agents in the Werewolf Game with Iterative Latent Space Policy Optimization
• Authors: Zelai Xu, Wanjun Gu, Chao Yu, Yi Wu, Yu Wang,
• Abstract summary: We propose Latent Space Policy Optimization (LSPO), an iterative framework that combines game-theoretic methods with fine-tuning to build strategic language agents.<n>Experiment on the Werewolf game shows that our agents iteratively expand the strategy space with improving performance and outperform existing Werewolf agents.
• Score: 13.496120603859701
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language model (LLM) agents have recently demonstrated impressive capabilities in various domains like open-ended conversation and multi-step decision-making. However, it remains challenging for these agents to solve strategic language games, such as Werewolf, which demand both strategic decision-making and free-form language interactions. Existing LLM agents often suffer from intrinsic bias in their action distributions and limited exploration of the unbounded text action space, resulting in suboptimal performance. To address these challenges, we propose Latent Space Policy Optimization (LSPO), an iterative framework that combines game-theoretic methods with LLM fine-tuning to build strategic language agents. LSPO leverages the observation that while the language space is combinatorially large, the underlying strategy space is relatively compact. We first map free-form utterances into a finite latent strategy space, yielding an abstracted extensive-form game. Then we apply game-theoretic methods like Counterfactual Regret Minimization (CFR) to optimize the policy in the latent space. Finally, we fine-tune the LLM via Direct Preference Optimization (DPO) to align with the learned policy. By iteratively alternating between these steps, our LSPO agents progressively enhance both strategic reasoning and language communication. Experiment on the Werewolf game shows that our agents iteratively expand the strategy space with improving performance and outperform existing Werewolf agents, underscoring their effectiveness in free-form language games with strategic interactions.

Related papers

• Navigating Motion Agents in Dynamic and Cluttered Environments through LLM Reasoning [69.5875073447454]
  This paper advances motion agents empowered by large language models (LLMs) toward autonomous navigation in dynamic and cluttered environments.<n>Our training-free framework supports multi-agent coordination, closed-loop replanning, and dynamic obstacle avoidance without retraining or fine-tuning.
  arXiv  Detail & Related papers  (2025-03-10T13:39:09Z)
• Simulation of Language Evolution under Regulated Social Media Platforms: A Synergistic Approach of Large Language Models and Genetic Algorithms [6.550725258692423]
  Social media platforms frequently impose restrictive policies to moderate user content, prompting the emergence of creative evasion language strategies. This paper presents a multi-agent framework based on Large Language Models (LLMs) to simulate the iterative evolution of language strategies under regulatory constraints.
  arXiv  Detail & Related papers  (2025-02-26T14:59:27Z)
• Policy Learning with a Natural Language Action Space: A Causal Approach [24.096991077437146]
  This paper introduces a novel causal framework for multi-stage decision-making in natural language action spaces. Our approach employs Q-learning to estimate Dynamic Treatment Regimes (DTR) through a single model. A key technical contribution of our approach is a decoding strategy that translates optimized embeddings back into coherent natural language.
  arXiv  Detail & Related papers  (2025-02-24T17:26:07Z)
• EPO: Explicit Policy Optimization for Strategic Reasoning in LLMs via Reinforcement Learning [69.55982246413046]
  We propose explicit policy optimization (EPO) for strategic reasoning.<n>We train the strategic reasoning model via multi-turn reinforcement learning (RL),utilizing process rewards and iterative self-play.<n>Our findings reveal various collaborative reasoning mechanisms emergent in EPO and its effectiveness in generating novel strategies.
  arXiv  Detail & Related papers  (2025-02-18T03:15:55Z)
• Lens: Rethinking Multilingual Enhancement for Large Language Models [70.85065197789639]
  Lens is a novel approach to enhance multilingual capabilities of large language models (LLMs) It operates by manipulating the hidden representations within the language-agnostic and language-specific subspaces from top layers of LLMs. It achieves superior results with much fewer computational resources compared to existing post-training approaches.
  arXiv  Detail & Related papers  (2024-10-06T08:51:30Z)
• 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)
• STRIDE: A Tool-Assisted LLM Agent Framework for Strategic and Interactive Decision-Making [43.734386326024016]
  Large Language Models (LLMs) have revolutionized natural language processing, showing remarkable linguistic proficiency and reasoning capabilities. This paper presents a novel framework equipped with memory and specialized tools to enhance their strategic decision-making capabilities.
  arXiv  Detail & Related papers  (2024-05-25T23:25:10Z)
• Language Evolution for Evading Social Media Regulation via LLM-based Multi-agent Simulation [6.550725258692423]
  Social media platforms such as Twitter, Reddit, and Sina Weibo play a crucial role in global communication but often encounter strict regulations in geopolitically sensitive regions. This paper proposes a multi-agent simulation framework using Large Language Models (LLMs) to explore the evolution of user language in regulated social media environments.
  arXiv  Detail & Related papers  (2024-05-05T09:02:54Z)
• How Can LLM Guide RL? A Value-Based Approach [68.55316627400683]
  Reinforcement learning (RL) has become the de facto standard practice for sequential decision-making problems by improving future acting policies with feedback. Recent developments in large language models (LLMs) have showcased impressive capabilities in language understanding and generation, yet they fall short in exploration and self-improvement capabilities. We develop an algorithm named LINVIT that incorporates LLM guidance as a regularization factor in value-based RL, leading to significant reductions in the amount of data needed for learning.
  arXiv  Detail & Related papers  (2024-02-25T20:07:13Z)
• Steering Language Models with Game-Theoretic Solvers [43.023261136434876]
  We introduce a framework that allows equilibrium solvers to work over the space of natural language dialogue generated by large language models (LLMs)<n>Specifically, by modelling the players, strategies and payoffs in a "game" of dialogue, we create a binding from natural language interactions to the conventional symbolic logic of game theory.<n>We focus on three domains that require different negotiation strategies: scheduling meetings, trading fruit and debate, and evaluate an LLM's generated language when guided by solvers.
  arXiv  Detail & Related papers  (2024-01-24T22:22:00Z)
• LMRL Gym: Benchmarks for Multi-Turn Reinforcement Learning with Language Models [56.25156596019168]
  This paper introduces the LMRL-Gym benchmark for evaluating multi-turn RL for large language models (LLMs) Our benchmark consists of 8 different language tasks, which require multiple rounds of language interaction and cover a range of tasks in open-ended dialogue and text games.
  arXiv  Detail & Related papers  (2023-11-30T03:59:31Z)
• Zero-Shot Goal-Directed Dialogue via RL on Imagined Conversations [70.7884839812069]
  Large language models (LLMs) have emerged as powerful and general solutions to many natural language tasks. However, many of the most important applications of language generation are interactive, where an agent has to talk to a person to reach a desired outcome. In this work, we explore a new method for adapting LLMs with RL for such goal-directed dialogue.
  arXiv  Detail & Related papers  (2023-11-09T18:45:16Z)
• ALYMPICS: LLM Agents Meet Game Theory -- Exploring Strategic Decision-Making with AI Agents [77.34720446306419]
  Alympics is a systematic simulation framework utilizing Large Language Model (LLM) agents for game theory research. Alympics creates a versatile platform for studying complex game theory problems.
  arXiv  Detail & Related papers  (2023-11-06T16:03:46Z)
• Leveraging Word Guessing Games to Assess the Intelligence of Large Language Models [105.39236338147715]
  The paper is inspired by the popular language game Who is Spy'' We develop DEEP to evaluate LLMs' expression and disguising abilities. We then introduce SpyGame, an interactive multi-agent framework.
  arXiv  Detail & Related papers  (2023-10-31T14:37:42Z)
• Language Agents with Reinforcement Learning for Strategic Play in the Werewolf Game [40.438765131992525]
  We develop strategic language agents that generate flexible language actions and possess strong decision-making abilities. To mitigate the intrinsic bias in language actions, our agents use an LLM to perform deductive reasoning and generate a diverse set of action candidates. Experiments show that our agents overcome the intrinsic bias and outperform existing LLM-based agents in the Werewolf game.
  arXiv  Detail & Related papers  (2023-10-29T09:02:57Z)
• Exploring Large Language Models for Communication Games: An Empirical Study on Werewolf [19.39740531672788]
  We propose a tuning-free framework to engage large language models in communication games. An empirical study on the representative and widely-studied communication game, Werewolf'', demonstrates that our framework can effectively play Werewolf game without tuning the parameters of the LLMs.
  arXiv  Detail & Related papers  (2023-09-09T01:56:40Z)
• Strategic Reasoning with Language Models [35.63300060111918]
  Strategic reasoning enables agents to cooperate, communicate, and compete with other agents in diverse situations. Existing approaches to solving strategic games rely on extensive training, yielding strategies that do not generalize to new scenarios or games without retraining. This paper introduces an approach that uses pretrained Large Language Models with few-shot chain-of-thought examples to enable strategic reasoning for AI agents.
  arXiv  Detail & Related papers  (2023-05-30T16:09:19Z)
• Offline Reinforcement Learning for Mixture-of-Expert Dialogue Management [36.254564021059515]
  Reinforcement learning (RL) has shown great promise for developing dialogue management (DM) agents that are non-myopic. We develop a variety of RL algorithms, specialized to dialogue planning, that leverage recent Mixture-of-Expert Language Models (MoE-LMs) By exploiting MoE-LM structure, our methods significantly reduce the size of the action space and improve the efficacy of RL-based DM.
  arXiv  Detail & Related papers  (2023-02-21T18:02:20Z)
• Is Reinforcement Learning (Not) for Natural Language Processing?: Benchmarks, Baselines, and Building Blocks for Natural Language Policy Optimization [73.74371798168642]
  We introduce an open-source modular library, RL4LMs, for optimizing language generators with reinforcement learning. Next, we present the GRUE benchmark, a set of 6 language generation tasks which are supervised not by target strings, but by reward functions. Finally, we introduce an easy-to-use, performant RL algorithm, NLPO, that learns to effectively reduce the action space in language generation.
  arXiv  Detail & Related papers  (2022-10-03T21:38:29Z)

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.