Search papers, labs, and topics across Lattice.
This paper introduces LifelongVLA, a novel framework for robotic manipulation that addresses the plasticity-stability trade-off in lifelong Vision-Language-Action learning. By employing a dual-timescale adaptation mechanism with a short-term adapter for plasticity and a long-term adapter for stability, the framework enables robots to efficiently learn new tasks while retaining previously acquired skills. Experimental results on an xArm robot reveal that LifelongVLA significantly outperforms existing models, showcasing effective skill expansion and robust retention of learned behaviors without the need for extensive retraining.
Robots can now learn new tasks in real-time while retaining past skills, thanks to a groundbreaking dual-timescale adaptation mechanism.
Similar to the natural capabilities of humans to sequentially learn new tasks, robots with Vision-Language-Action (VLA) models should possess lifelong learning ability to learn a new task when deployed in open-world environments. However, most recently proposed lifelong learning models aim to effectively learn the current task (plasticity) or maintain high accuracy on previous tasks (stability), while the plasticity-stability trade-off remains largely unsolved in robotic manipulation models. To address this fundamental challenge, we propose a cache-efficient lifelong Vision-Language-Action learning framework for robotic manipulation (i.e., LifelongVLA), which alleviates the plasticity-stability trade-off with a dual-timescale adaptation mechanism while achieving low-cost robotic deployment with a cache-efficient replay strategy. More concretely, we propose a dual-timescale LoRA gating module to decompose VLA adaptation into two lightweight pathways: a short-term adapter for plasticity and a long-term adapter for stable consolidation. These pathways are integrated via a task-aware gate, enabling explicit control of the plasticity-stability trade-off. In the skill replay phase, a cache-efficient stochastic replay strategy is proposed to preserve more balanced retention signals without full-trajectory storage. Finally, experiments show that LifelongVLA outperforms existing baselines, demonstrating efficient skill expansion, robust retention of learned manipulation behaviors, and reduced reliance on retraining for real-world deployment on an xArm robot.
