Search papers, labs, and topics across Lattice.
This paper introduces the Skill-Compositional Experts (SCE) framework for Embodied Continual Learning (ECL), addressing the challenge of catastrophic forgetting in robots learning new manipulation tasks. By utilizing Compositional Skill Grounding (CSG) to decompose task demonstrations into reusable skills, SCE enhances structured skill reuse and coherent task execution through its Dual Execution-and-Transition Experts (DETE) mechanism. Experimental results on LIBERO benchmarks and real-world tasks reveal that SCE significantly improves both retention of previously learned behaviors and overall task performance, while further analyses confirm its robustness against feature drift.
Skill-Compositional Experts can dramatically enhance a robot's ability to learn new tasks without forgetting old ones, tackling a critical barrier in embodied AI.
Embodied Continual Learning (ECL) aims to enable robots to continually acquire new manipulation tasks while retaining previously learned behaviors under closed-loop control. Compared with conventional continual learning, ECL suffers from more severe catastrophic forgetting. Feature drift accumulated under closed-loop control progressively propagates through sequential decision-making, leading to degradation of previously learned behaviors. A key challenge in ECL lies in structured skill reuse across continually evolving tasks, since existing methods primarily focus on skill learning without explicitly organizing them for coherent task execution. To address this issue, we propose SCE, a Skill-Compositional Experts framework for ECL. SCE builds a skill base via Compositional Skill Grounding (CSG), which decomposes task demonstrations into reusable skills. Based on this, Dual Execution-and-Transition Experts (DETE) enable new task learning through skill composition, where one branch ensures skill execution and the other supports transitions between skills for coherent behavior. Experiments on LIBERO benchmarks and real-world manipulation tasks demonstrate that SCE consistently improves retention and overall task performance. Further feature drift analyses and ablation studies verify the effectiveness of our method. Project website: https://eqcy.github.io/sce/.
