Non-Markovian Long-Horizon Robot Manipulation via Keyframe Chaining

TL;DR

This paper presents Keyframe-Chaining VLA, a novel framework that enhances long-horizon robot manipulation by explicitly modeling non-Markovian dependencies through keyframe selection and dynamic retrieval, improving task success rates.

Contribution

It introduces an automatic keyframe selector and progress-aware retrieval mechanism to explicitly incorporate long-term dependencies into vision-language-action models for robot manipulation.

Findings

Achieves superior performance on non-Markovian manipulation tasks.

Effectively models long-horizon dependencies with keyframe chaining.

Demonstrates improved task success rates over baseline methods.

Abstract

Existing Vision-Language-Action (VLA) models often struggle to generalize to long-horizon tasks due to their heavy reliance on immediate observations. While recent studies incorporate retrieval mechanisms or extend context windows to handle procedural tasks, they often struggle to capture Non-Markovian dependencies, where optimal actions rely solely on specific past states rather than the current observation. To address this, we introduce Keyframe-Chaining VLA, a framework that extracts and links key historical frames to model long-horizon dependencies. Specifically, we propose an automatic keyframe selector that learns a discriminative embedding space, effectively identifying distinct state transitions. To capture task-critical information, we design a progress-aware query mechanism that dynamically retrieves historical frames based on their temporal relevance to the current execution phase. These selected keyframes are integrated into the VLA as interleaved visual tokens, explicitly grounding the policy in the long-horizon temporal context. Finally, we introduce a suite of four Non-Markovian manipulation tasks built upon the ManiSkill simulator to measure task success rates. Experimental results demonstrate that our method achieves superior performance, effectively tackling robot manipulation tasks characterized by long-horizon temporal dependencies. Code is available at https://github.com/cytoplastm/KC-VLA.