用直觉模型增强世界模型，让像素级规划更可靠
IMWM: Intuition Models Complement World Models for Latent Planning
- 引入直觉模型从演示中识别好动作，与世界模型协同规划
- 在四类任务上成功率均提升，最高增益达28.5个百分点
- 适合需要高效决策的视觉控制场景，如机器人抓取
基于学习的潜在世界模型是实现从原始像素进行控制的有前景路径，但仅靠强大的世界模型仍不够。实验表明：即使使用理想化的环境动态滚动生成（替代学习的前向预测器），有限预算的采样规划器在某些任务上仍会失败，说明瓶颈可能在于搜索而非世界模型精度。为此，我们提出 IMWM（直觉模型 + 世界模型），将世界模型与从示范中训练的直觉模型结合，通过三个轻量组件协同：(i) 检索初始化，从检索到的示范中初始化动作提议；(ii) 混合代价，融合直觉得分与世界模型滚动代价；(iii) 可靠性门控，动态调节规划器对直觉的信任程度。在四个基于像素的目标达成任务（Two-Room、Reacher、Push-T、OGBench-Cube）上，IMWM 的平均成功率均高于仅用世界模型的规划器，其中 Two-Room 提升 11.5 个百分点至 99.2%，OGBench-Cube 提升 28.5 个百分点至 94.7%。
原文摘要 · Abstract (English)
Planning with a learned latent world model is a promising route to control from raw pixels, but a strong world model alone is not enough. We show this experimentally: even with a perfect world model (operationalized by replacing the learned forward predictor with an idealized rollout of the true environment dynamics), a finite-budget sample-based planner still fails on some tasks, indicating that the bottleneck can lie in search rather than in world-model accuracy. Motivated by this gap, we propose IMWM (Intuition Model + World Model), which pairs the world model with an intuition model trained from demonstrations to recognize promising actions. The two models collaborate through three lightweight components: (i) Retrieval Initialization, which initializes the planner's action proposal from a retrieved demonstration; (ii) Hybrid Cost, which combines the intuition score with the world-model rollout cost; and (iii) a Reliability Gate, which adjusts how much the planner trusts intuition in each setting. Across four pixel-based goal-reaching tasks (Two-Room, Reacher, Push-T, and OGBench-Cube), IMWM has higher mean success than the world-model-only planner on all four, with the largest gains on Two-Room (99.2%, +11.5 percentage points) and OGBench-Cube (94.7%, +28.5 percentage points).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。