Dyna-Mind: Learning to Simulate from Experience for Better AI Agents

• URL: http://arxiv.org/abs/2510.09577v1
• Date: Fri, 10 Oct 2025 17:30:18 GMT
• Title: Dyna-Mind: Learning to Simulate from Experience for Better AI Agents
• Authors: Xiao Yu, Baolin Peng, Michel Galley, Hao Cheng, Qianhui Wu, Janardhan Kulkarni, Suman Nath, Zhou Yu, Jianfeng Gao,
• Abstract summary: We argue that current AI agents need ''vicarious trial and error'' - the capacity to mentally simulate alternative futures before acting.<n>We introduce Dyna-Mind, a two-stage training framework that explicitly teaches (V)LM agents to integrate such simulation into their reasoning.
• Score: 62.21219817256246
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reasoning models have recently shown remarkable progress in domains such as math and coding. However, their expert-level abilities in math and coding contrast sharply with their performance in long-horizon, interactive tasks such as web navigation and computer/phone-use. Inspired by literature on human cognition, we argue that current AI agents need ''vicarious trial and error'' - the capacity to mentally simulate alternative futures before acting - in order to enhance their understanding and performance in complex interactive environments. We introduce Dyna-Mind, a two-stage training framework that explicitly teaches (V)LM agents to integrate such simulation into their reasoning. In stage 1, we introduce Reasoning with Simulations (ReSim), which trains the agent to generate structured reasoning traces from expanded search trees built from real experience gathered through environment interactions. ReSim thus grounds the agent's reasoning in faithful world dynamics and equips it with the ability to anticipate future states in its reasoning. In stage 2, we propose Dyna-GRPO, an online reinforcement learning method to further strengthen the agent's simulation and decision-making ability by using both outcome rewards and intermediate states as feedback from real rollouts. Experiments on two synthetic benchmarks (Sokoban and ALFWorld) and one realistic benchmark (AndroidWorld) demonstrate that (1) ReSim effectively infuses simulation ability into AI agents, and (2) Dyna-GRPO leverages outcome and interaction-level signals to learn better policies for long-horizon, planning-intensive tasks. Together, these results highlight the central role of simulation in enabling AI agents to reason, plan, and act more effectively in the ever more challenging environments.

Related papers

• TongSIM: A General Platform for Simulating Intelligent Machines [59.27575233453533]
  Embodied intelligence focuses on training agents within realistic simulated environments.<n>TongSIM is a high-fidelity, general-purpose platform for training and evaluating embodied agents.
  arXiv  Detail & Related papers  (2025-12-23T10:00:43Z)
• Scaling Agent Learning via Experience Synthesis [100.42712232390532]
  Reinforcement learning can empower autonomous agents by enabling self-improvement through interaction.<n>But its practical adoption remains challenging due to costly rollouts, limited task diversity, unreliable reward signals, and infrastructure complexity.<n>We introduce DreamGym, the first unified framework designed to synthesize diverse experiences with scalability in mind.
  arXiv  Detail & Related papers  (2025-11-05T18:58:48Z)
• See, Think, Act: Online Shopper Behavior Simulation with VLM Agents [58.92444959954643]
  This paper investigates the integration of visual information, specifically webpage screenshots, into behavior simulation via VLMs.<n>We employ SFT for joint action prediction and rationale generation, conditioning on the full interaction context.<n>To further enhance reasoning capabilities, we integrate RL with a hierarchical reward structure, scaled by a difficulty-aware factor.
  arXiv  Detail & Related papers  (2025-10-22T05:07:14Z)
• Dyna-Think: Synergizing Reasoning, Acting, and World Model Simulation in AI Agents [57.35214204211501]
  We propose Dyna-Think, a thinking framework that integrates planning with an internal world model with reasoning and acting to enhance AI agent performance.<n>DIT reconstructs the thinking process of R1 to focus on performing world model simulation relevant to the proposed (and planned) action, and trains the policy using this reconstructed data.<n>DDT uses a two-stage training process to first improve the agent's world modeling ability via objectives such as state prediction or critique generation, and then improve the agent's action via policy training.
  arXiv  Detail & Related papers  (2025-05-31T00:10:18Z)
• YuLan-OneSim: Towards the Next Generation of Social Simulator with Large Language Models [50.35333054932747]
  We introduce a novel social simulator called YuLan-OneSim.<n>Users can simply describe and refine their simulation scenarios through natural language interactions with our simulator.<n>We implement 50 default simulation scenarios spanning 8 domains, including economics, sociology, politics, psychology, organization, demographics, law, and communication.
  arXiv  Detail & Related papers  (2025-05-12T14:05:17Z)
• AI Metropolis: Scaling Large Language Model-based Multi-Agent Simulation with Out-of-order Execution [15.596642151634319]
  AI Metropolis is a simulation engine that improves the efficiency of LLM agent simulations by incorporating out-of-order execution scheduling. Our evaluations demonstrate that AI Metropolis achieves speedups from 1.3x to 4.15x over standard parallel simulation with global synchronization.
  arXiv  Detail & Related papers  (2024-11-05T21:54:14Z)
• A Platform-Agnostic Deep Reinforcement Learning Framework for Effective Sim2Real Transfer towards Autonomous Driving [0.0]
  Deep Reinforcement Learning (DRL) has shown remarkable success in solving complex tasks. transferring DRL agents to the real world is still challenging due to the significant discrepancies between simulation and reality. We propose a robust DRL framework that leverages platform-dependent perception modules to extract task-relevant information.
  arXiv  Detail & Related papers  (2023-04-14T07:55:07Z)
• RoboTHOR: An Open Simulation-to-Real Embodied AI Platform [56.50243383294621]
  We introduce RoboTHOR to democratize research in interactive and embodied visual AI. We show there exists a significant gap between the performance of models trained in simulation when they are tested in both simulations and their carefully constructed physical analogs.
  arXiv  Detail & Related papers  (2020-04-14T20:52:49Z)

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.