联邦学习中的时间漂移——通过经验回放缓解灾难性遗忘
Federated Learning Under Temporal Drift -- Mitigating Catastrophic Forgetting via Experience Replay
AI总结:
本研究通过客户端侧经验回放方法缓解联邦学习中的时间漂移问题,实验表明该方法能有效恢复模型性能并防止灾难性遗忘。
摘要:
联邦学习在时间概念漂移下表现不佳,其中客户端数据分布随时间变化。我们证明,在Fashion-MNIST上,标准FedAvg在季节性漂移下遭受灾难性遗忘,准确率从74%下降到28%。我们提出客户端侧经验回放,其中每个客户端在本地训练期间维护一个包含过去样本和当前数据的较小缓冲区。这种简单的方法无需更改服务器聚合。实验显示,每个类别50个样本的缓冲区可将性能恢复到78-82%,有效防止遗忘。我们的消融研究揭示了随着缓冲区大小增加,记忆与准确率之间的明显权衡。
英文摘要:
Federated Learning struggles under temporal concept drift where client data distributions shift over time. We demonstrate that standard FedAvg suffers catastrophic forgetting under seasonal drift on Fashion-MNIST, with accuracy dropping from 74% to 28%. We propose client-side experience replay, where each client maintains a small buffer of past samples mixed with current data during local training. This simple approach requires no changes to server aggregation. Experiments show that a 50-sample-per-class buffer restores performance to 78-82%, effectively preventing forgetting. Our ablation study reveals a clear memory-accuracy trade-off as buffer size increases.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。