ALAS: Adaptive Long-Horizon Action Synthesis via Async-pathway Stream Disentanglement

TL;DR

ALAS introduces a biologically inspired dual-stream framework for long-horizon tasks in human-scene interaction, enabling better generalization and efficiency across domains and skills.

Contribution

This work proposes ALAS, a novel cross-domain learning framework with environment and skill disentanglement inspired by brain pathways, improving long-horizon task performance.

Findings

ALAS achieves 23% higher subtask success rate on average.

ALAS improves execution efficiency by 29%.

Extensive experiments validate ALAS's cross-domain and cross-skill transfer capabilities.

Abstract

Long-Horizon (LH) tasks in Human-Scene Interaction (HSI) are complex multi-step tasks that require continuous planning, sequential decision-making, and extended execution across domains to achieve the final goal. However, existing methods heavily rely on skill chaining by concatenating pre-trained subtasks, with environment observations and self-state tightly coupled, lacking the ability to generalize to new combinations of environments and skills, failing to complete various LH tasks across domains. To solve this problem, this paper presents ALAS, a cross-domain learning framework for LH tasks via biologically inspired dual-stream disentanglement. Inspired by the brain's "where-what" dual pathway mechanism, ALAS comprises two core modules: i) an environment learning module for spatial understanding, which captures object functions, spatial relationships, and scene semantics, achieving cross-domain transfer through complete environment-self disentanglement; ii) a skill learning module for task execution, which processes self-state information including joint degrees of freedom and motor patterns, enabling cross-skill transfer through independent motor pattern encoding. We conducted extensive experiments on various LH tasks in HSI scenes. Compared with existing methods, ALAS can achieve an average subtasks success rate improvement of 23\% and average execution efficiency improvement of 29\%.