通过推理时扩散缩放实现稳定长时程视觉规划
Compositional Visual Planning via Inference-Time Diffusion Scaling
- 基于重叠视频片段的链式因子图，利用预训练模型提供局部先验
- 推理时通过同步异步消息传递确保边界一致性，提升全局规划稳定性
- 无需训练即可泛化到未见过的起止组合，适合长程机器人任务
扩散模型在短时程机器人规划中表现优异，但扩展到长时程任务仍受计算约束和训练数据有限的限制。现有组合方法通过分别去噪各组件并平均重叠区域来拼接短段，但在噪声数据空间中因因子分解假设失效，导致全局计划不一致。本文提出：稳定组合生成的关键在于对估计的干净数据（Tweedie估计）施加边界一致性，而非对噪声中间状态进行约束。方法将长时程规划建模为一系列重叠视频块构成的链式因子图，由预训练的短时程视频扩散模型提供局部先验。推理时，通过一种新型的同步与异步消息传递结合机制，在Tweedie估计上强制边界一致，从而生成全局一致的引导信号，且无需额外训练。该免训练框架显著优于现有基线，能有效泛化至原始训练数据中未出现的起止组合。
原文摘要 · Abstract (English)
Diffusion models excel at short-horizon robot planning, yet scaling them to long-horizon tasks remains challenging due to computational constraints and limited training data. Existing compositional approaches stitch together short segments by separately denoising each component and averaging overlapping regions. However, this suffers from instability as the factorization assumption breaks down in noisy data space, leading to inconsistent global plans. We propose that the key to stable compositional generation lies in enforcing boundary agreement on the estimated clean data (Tweedie estimates) rather than on noisy intermediate states. Our method formulates long-horizon planning as inference over a chain-structured factor graph of overlapping video chunks, where pretrained short-horizon video diffusion models provide local priors. At inference time, we enforce boundary agreement through a novel combination of synchronous and asynchronous message passing that operates on Tweedie estimates, producing globally consistent guidance without requiring additional training. Our training-free framework demonstrates significant improvements over existing baselines, effectively generalizing to unseen start-goal combinations that were not present in the original training data. Project website: https://comp-visual-planning.github.io/
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。