当生成器回放退化时：面向异构联邦类增量学习的投影排练编排
When Generator Replay Degrades: Projected Rehearsal Orchestration for Heterogeneous Federated Class-Incremental Learning
AI总结：
针对异构联邦类增量学习中客户端标签子集不同、任务阶段不一致导致的旧知识遗忘问题，提出投影排练编排框架PRO及增强版PRO-MAX，通过服务器端维护紧凑类级投影记忆并实现平衡伪多任务训练，在图像、文本和图基准上提升异构流下的保留与最终效用。
摘要：
联邦类增量学习（FCIL）在客户端观察到不同标签子集、在不同阶段推进任务以及为相同语义概念提供不均匀监督时变得极其困难。现有的FCIL方法通常通过输入空间合成来保留旧知识，但在异构任务流下可能脆弱且难以跨模态迁移。为缓解这些问题，我们提出PRO，一个用投影排练编排替代合成输入回放的框架。为去除外部预训练，我们在相同的预热条件下评估所有方法。此后，PRO在服务器上维护紧凑的类级投影记忆，并允许客户端在当前示例和旧投影记忆上执行平衡的伪多任务训练。为处理更强的表示漂移，我们进一步引入PRO-MAX，它在保持相同服务器轻量原则（服务器仅聚合模型更新和记忆统计）的同时，用邻域加权记忆对齐增强PRO。在图像、文本和图基准上，PRO和PRO-MAX在异构流下提高了保留和最终效用，同时在同构FCIL中保持竞争力。即使基线获得更大的回放预算，它们在监督不平衡和阶段错位下也会退化，表明仅靠回放数量无法解决回放质量失败。额外的弱任务诊断进一步表明，更大的回放不匹配与更大的下游退化相关，而我们的方法使投影记忆与不断演化的表示保持更好对齐。
英文摘要：
Federated class-incremental learning (FCIL) becomes substantially harder when clients observe different label subsets, progress through tasks at different stages, and provide uneven supervision for the same semantic concepts. Existing FCIL methods often preserve old knowledge through input-space synthesis, but they can be fragile under heterogeneous task streams and difficult to transfer across modalities. To alleviate such issues, we propose PRO, a framework that replaces synthetic input replay with projected rehearsal orchestration. To remove external pretraining, we evaluate all methods under the same warmup. After this, PRO maintains compact class-level projected memories on the server and allows clients perform balanced pseudo multi-task training over current examples and old projected memories. To handle stronger representation drift, we further introduce PRO-MAX, which augments PRO with neighborhood-weighted memory alignment while preserving the same server-light principle that the server only aggregates model updates and memory statistics. Across image, text, and graph benchmarks, PRO and PRO-MAX improve retention and final utility under heterogeneous streams while remaining competitive in homogeneous FCIL. Even when baselines are given expanded replay budgets, they degrade under supervision imbalance and stage misalignment, indicating that replay quantity alone does not resolve replay-quality failures. Additional weak-task diagnostics further show that larger replay mismatch is associated with larger downstream degradation, while our method keeps projected memories better aligned with the evolving representation.
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。
