首个面向机器人操作的联邦学习基准，支持隐私保护下的分布式训练
FLAME: A Federated Learning Benchmark for Robotic Manipulation
- 构建了16万条专家示范数据集，覆盖多种模拟环境
- 验证了标准联邦学习算法在分布式策略学习中的可行性
- 适合关注隐私保护与可扩展性研究的机器人学者
近期机器人操作进展得益于大规模跨环境数据集。传统上，这些数据集采用集中式方式训练策略，引发可扩展性、适应性与数据隐私问题。尽管联邦学习能实现去中心化、隐私保护的训练，但其在机器人操作领域的应用仍不充分。本文提出FLAME（Federated Learning Across Manipulation Environments），首个专为机器人操作设计的联邦学习基准。FLAME包含：(i) 超过16万条多任务操作的专家示范数据，采集自广泛模拟环境；(ii) 支持联邦环境下机器人策略学习的训练与评估框架。我们对标准联邦学习算法进行了评估，验证其在分布式策略学习中的潜力，并揭示关键挑战。该基准为可扩展、自适应、隐私友好的机器人学习奠定了基础。
原文摘要 · Abstract (English)
Recent progress in robotic manipulation has been fueled by large-scale datasets collected across diverse environments. Training robotic manipulation policies on these datasets is traditionally performed in a centralized manner, raising concerns regarding scalability, adaptability, and data privacy. While federated learning enables decentralized, privacy-preserving training, its application to robotic manipulation remains largely unexplored. We introduce FLAME (Federated Learning Across Manipulation Environments), the first benchmark designed for federated learning in robotic manipulation. FLAME consists of: (i) a set of large-scale datasets of over 160,000 expert demonstrations of multiple manipulation tasks, collected across a wide range of simulated environments; (ii) a training and evaluation framework for robotic policy learning in a federated setting. We evaluate standard federated learning algorithms in FLAME, showing their potential for distributed policy learning and highlighting key challenges. Our benchmark establishes a foundation for scalable, adaptive, and privacy-aware robotic learning.
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。