面向联邦持续学习的参数高效模块闭式合并
Closed-form merging of parameter-efficient modules for Federated Continual Learning
AI总结：
本文针对联邦类增量学习中的模型合并问题，提出交替优化策略 LoRM，通过逐矩阵求解 LoRA 参数实现闭式唯一解，并在多个 FCIL 场景中取得最先进性能。
摘要：
模型合并已成为深度学习中的一项关键技术，它能够在保持性能和可扩展性的同时，将多个模型整合为一个统一系统。在这方面，低秩适配技术（如 LoRA）的组合性质已被证明是有益的，因为简单地对 LoRA 模块求平均即可得到一个能在很大程度上整合所有单个模块能力的单一模型。在 LoRA 的基础上，我们更进一步，要求合并后的模型能够匹配所有已学习模块的响应。以闭式形式求解该目标会得到一个以 A 和 B 为未知变量的不确定系统，这表明存在无穷多个闭式解。为应对这一挑战，我们提出了 LoRM，这是一种交替优化策略，每次只训练一个 LoRA 矩阵。这使得可以分别求解每个未知变量，从而找到唯一解。我们将所提出的方法应用于联邦类增量学习（FCIL），确保模型响应在客户端之间以及跨任务之间保持一致。我们的方法在一系列 FCIL 场景中展现了最先进的性能。用于复现实验的代码可在 github.com/aimagelab/fed-mammoth 获取。
英文摘要：
Model merging has emerged as a crucial technique in Deep Learning, enabling the integration of multiple models into a unified system while preserving perfor-mance and scalability. In this respect, the compositional properties of low-rank adaptation techniques (e.g., LoRA) have proven beneficial, as simple averaging LoRA modules yields a single model that mostly integrates the capabilities of all individual modules. Building on LoRA, we take a step further by imposing that the merged model matches the responses of all learned modules. Solving this objective in closed form yields an indeterminate system with A and B as unknown variables, indicating the existence of infinitely many closed-form solutions. To address this challenge, we introduce LoRM, an alternating optimization strategy that trains one LoRA matrix at a time. This allows solving for each unknown variable individually, thus finding a unique solution. We apply our proposed methodology to Federated Class-Incremental Learning (FCIL), ensuring alignment of model responses both between clients and across tasks. Our method demonstrates state-of-the-art performance across a range of FCIL scenarios. The code to reproduce our experiments is available at github.com/aimagelab/fed-mammoth.
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。
