Continual Robot Skill and Task Learning via Dialogue

View PDF HTML (experimental)

Abstract:Interactive robot learning is a challenging problem as the robot is present with human users who expect the robot to learn novel skills to solve novel tasks perpetually with sample efficiency. In this work we present a framework for robots to continually learn tasks and visuo-motor skills and query for novel skills via dialog interactions with human users. Our robot agent maintains a skill library, and uses an existing LLM to perform grounded dialog interactions to query unknown skills from real human users. We developed a novel visual-motor control policy Action Chunking Transformer with Low Rank Adaptation (ACT-LoRA) that can continually learn novel skills using only a few demonstrations which is critical in human-robot interaction scenarios. The paper has twin goals: Firstly to demonstrate better continual learning in simulation; and secondly, to demonstrate the use of our dialog based learning framework in a realistic human-robot interaction use case. Our ACT-LoRA policy consistently outperforms a GMM-LoRA baseline on multiple continual learning simulation benchmarks by achieving > 300% improvements on novel skills, while achieving comparable performance in existing skills. Moreover, with our IRB approved human-subjects study we demonstrate that our dialog based continual learning framework allows users to teach robots cooking skills successfully (100%) while spending a higher ratio of time on finishing an auxiliary distraction tasks in the test phase of the study compared to a non-learning language based agent (p < 0.001).

Subjects:

Robotics (cs.RO); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)

Cite as: arXiv:2409.03166 [cs.RO]

(or arXiv:2409.03166v3 [cs.RO] for this version)

https://doi.org/10.48550/arXiv.2409.03166

arXiv-issued DOI via DataCite

Submission history

From: Weiwei Gu [view email] [v1] Thu, 5 Sep 2024 01:51:54 UTC (4,008 KB) [v2] Wed, 11 Sep 2024 21:52:22 UTC (4,009 KB) [v3] Sat, 28 Mar 2026 18:34:40 UTC (3,866 KB)