外观
论文arXiv 2610.1912915 天前
使用 Geometric Horizon Models 想象目标以实现策略泛化
原文链接：Imagining Goals with Geometric Horizon Models for Policy Generalization（arXiv 2610.19129 · 发布日期：2026-10-13）
Imagining Goals with Geometric Horizon Models for Policy Generalization
该工作面向零样本目标泛化，提出 TD-Aug 从 geometric horizon model 采样想象新目标，并在 stitching-based OGBench 任务验证。
核心要点
- 离线训练的通用策略虽可零样本或微调应用，但受建模目标和静态数据限制，泛化仍不足。
- TD-Aug 从 geometric horizon model 采样，直接想象由已有知识组合可达的新结果。
- 将想象出的未来结果作为目标训练 goal-conditioned BC 策略，以提升组合式目标泛化。
- 在 stitching-based OGBench 任务中，该训练显著改善了对未见目标任务的泛化。
验证范围与局限：验证范围限于摘要所述的 stitching-based OGBench 任务。
英文摘要 (Abstract)
Training from offline data has allowed for substantial progress in domains such as robotics, leading to general-purpose policies that can be easily applied zero-shot or efficiently finetuned for downstream tasks. However, these policies can still have poor generalization, both due to the choice of modeling objective and from learning from a static dataset. In this work, we focus on the challenging task of zero-shot goal generalization, where a policy is evaluated on unseen tasks that require combining its existing knowledge (compositional generalization). An avenue for improving a policy's generalization is by generating new experience with world models; however, such generation has proven difficult for longer horizons. Thus, to alleviate this issue, we propose TD-Aug, which samples from a geometric horizon model and allows for directly imagining novel outcomes that can be achieved by composing existing knowledge. We demonstrate that training on these future outcomes as goals for goal-conditioned BC policies significantly improves generalization in stitching-based OGBench tasks.
作者：Daniel Lawson, Dhruv Sreenivas, Adriana Hugessen, Khimya Khetarpal, Glen Berseth
来源
本页事实对应的公开页面，以原文为准。