用在线元学习适配器，让机器人持续学习新任务时更快更准。
Efficient Continual Adaptation of Pretrained Robotic Policy with Online Meta-Learned Adapters
- 通过元学习机制动态生成适配器，实现跨任务知识迁移。
- 在模拟和真实环境中，适应速度和效果均优于基线方法。
- 适合需要长期自主学习的机器人系统研究者使用。
持续适应对通用自主智能体至关重要。例如，经过多种技能预训练的家庭机器人仍需适应每个家庭的特定任务。受语言模型中参数高效微调的启发，先前工作探索了轻量级适配器来适应预训练策略，能保留预训练阶段学到的特征并表现出良好的适应性能。然而，这些方法将任务学习独立处理，限制了任务间的知识迁移。本文提出在线元学习适配器（OMLA）。不同于直接应用适配器，OMLA通过新颖的元学习目标，促进以往任务知识向当前任务的迁移。在模拟与真实环境中的大量实验表明，相较于基线方法，OMLA可实现更优的适应性能。
原文摘要 · Abstract (English)
Continual adaptation is essential for general autonomous agents. For example, a household robot pretrained with a repertoire of skills must still adapt to unseen tasks specific to each household. Motivated by this, building upon parameter-efficient fine-tuning in language models, prior works have explored lightweight adapters to adapt pretrained policies, which can preserve learned features from the pretraining phase and demonstrate good adaptation performances. However, these approaches treat task learning separately, limiting knowledge transfer between tasks. In this paper, we propose Online Meta-Learned adapters (OMLA). Instead of applying adapters directly, OMLA can facilitate knowledge transfer from previously learned tasks to current learning tasks through a novel meta-learning objective. Extensive experiments in both simulated and real-world environments demonstrate that OMLA can lead to better adaptation performances compared to the baseline methods. The project link: https://ricky-zhu.github.io/OMLA/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。