MemWM：记忆增强的基于文本的世界模型

中文摘要 AI 辅助

世界模型越来越多地用于智能体规划，通过预测环境状态如何响应智能体动作而演变来提供支持。然而流畅的下一状态预测仍可能遗漏任务关键事实、损坏产品属性或应用错误的转移规则。为解决此类系统性预测错误，我们提出MemWM，一种记忆增强的基于文本的世界模型。MemWM利用世界记忆（由转移规则、状态缓存和难以预测的事实组成的精选记忆库）来调控下一状态的想象。我们通过结构化状态保真度（SSF）评估事实状态保留情况，SSF通过特定于基准的事实和字段对预测状态进行评分。与SFT相比，记忆增强训练使SSF提升高达206.3%。在完整规划设置中，我们保持策略模型冻结，并提供策略侧世界技能：检索到的任务级技能和用于动作选择的逐步纠正指导。在ALFWorld、WebShop和ScienceWorld上，记忆增强智能体比经SFT训练的世界模型智能体提升了下游成功率，相对增益高达65.4%。敏感性分析进一步表明，在不同的记忆和动作预算设置下，检索到的记忆可提升任务成功率和效率。

英文摘要

World models are increasingly used to support planning in agents by predicting how environment states evolve in response to agent actions. Yet fluent next-state predictions can still omit task-critical facts, corrupt product attributes, or apply incorrect transition rules. To address such systematic prediction errors, we introduce MemWM, a memory-augmented text-based world model. MemWM uses world memory, a curated memory bank of transition rules, state caches, and hard-to-predict facts, to condition next-state imagination. We evaluate factual state preservation with Structured State Fidelity (SSF), which scores predicted states through benchmark-specific facts and fields. Compared with SFT, memory-augmented training improves SSF by up to 206.3%. In the full planning setting, we keep the policy model frozen and provide policy-side world skill: retrieved task-level skills and step-wise corrective guidance for action selection. Across ALFWorld, WebShop, and ScienceWorld, memory-augmented agents improve downstream success over an SFT-trained world-model agent, with up to a 65.4% relative gain. Sensitivity analyses further show that retrieved memory improves task success and efficiency under different memory and action-budget settings.

发表机构

• LMU Munich（慕尼黑大学）
• Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）
• Huawei Heisenberg Research Center（华为海森堡研究中心）
• Zhejiang University（浙江大学）
• Technical University of Munich (TUM)（慕尼黑工业大学）
• TU Berlin（柏林工业大学）
• Kiel University（基尔大学）

机构由 AI 辅助整理，请以论文原文为准。

补充信息

备注

Accepted to EMNLP 2026 (Main Conference)