生涯ロボット学習のための疎なアダプタ割当・再利用・統合手法SARAM
SARAM: Sparse Adapter Allocation with Reuse and Merging for Lifelong Robot Learning
新タスクごとに疎な層だけにアダプタを追加し既存の凍結アダプタを再利用、展開時はタスク署名で更新をバックボーンに統合することで、記憶容量と推論コストをタスク数に依存させずに生涯学習を実現した。
詳しい要約
1. どんなもの?
2. 先行研究と比べてどこがすごい?
3. 技術・手法の肝は?
4. どうやって有効だと検証した?
5. 議論はある?
6. 次に読むべき論文は?
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: SeongJun Jeong, Chungwoo Lee, Gi-Cheon Kang, Byoung-Tak Zhang
分類: cs.RO
原文アブストラクト
Lifelong robot learning requires acquiring new tasks sequentially without forgetting earlier ones. With large pre-trained vision-language-action (VLA) backbones, adapter-based methods effectively mitigate this challenge. However, their task-specific adapters and selection modules accumulate with the task stream, creating a scalability challenge in on-disk storage and deployment cost under fixed onboard resource budgets. We propose SARAM, a lifelong learning framework that limits storage growth while keeping GPU memory and per-step inference cost independent of the task stream length. For each new task, SARAM reuses frozen adapters from previous tasks and introduces new adapters only at sparsely selected layers, with a duplication penalty encouraging them to capture information not already represented by the adapter library. At deployment, a compact signature derived from backbone activation statistics identifies the task once per episode, and the corresponding task update is merged into the backbone, leaving only the backbone in the control loop. Across four LIBERO suites, SARAM matches or exceeds adapter-based baselines while adding up to 4.5x less on-disk state per task. GPU memory usage and control frequency remain comparable to those of the backbone as the task stream grows. Experiments on a real UR5 arm with a different backbone further demonstrate retention throughout a sequential task stream.
