AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control· 针对反事实模型预测控制的动作区分世界模型
Latent world models are typically trained to predict factual transitions, whereas model predictive control (MPC) must compare alternative actions from the same state. A model can therefore achieve low factual prediction error yet poorly distinguish candidate actions. We introduce AD-WM, an action-discriminative joint-embedding world model for counterfactual MPC. AD-WM combines residual latent dynamics with predictor-level action-recovery regularization, using inverse dynamics and a normalized recovery objective motivated by conditional mutual information. Both objectives encourage planning transitions to preserve action information; their auxiliary heads are discarded at test time, leaving MPC unchanged. On OGBench-Cube, AD-WM improves hard-start success from 3.7% to 52.0% over a matched L
提出 AD-WM 模型,增强反事实模型预测控制中的动作区分能力。
- 核心方法
- 通过联合嵌入世界模型,结合残留潜在动态和预测级别的动作恢复正则化(使用逆向动力学和归一化恢复目标),保留动作信息,提高动作区分度。
- 适合谁读
- 研究者、工程师
- 要解决的问题
- 现有世界模型在预测实际状态转换时表现良好,但在反事实模型预测控制中难以有效区分动作选项。
- 关键实验
- 在 OGBench-Cube 数据集上的实验结果显示,AD-WM 显著提高了任务成功率。
- 主要贡献
- AD-WM 模型在 OGBench-Cube 测试中大幅提高从困难起点开始的成功率,从 3.7% 提升至 52.0%。
- 意义与局限
- 该模型为提高反事实模型预测控制中的动作区分提供了有效方法,但辅助头部在测试时被丢弃,可能影响模型的泛化能力。