arXiv cs.MA· Yara Zgheib, Marc Antonini, Roula Nassif·· 9 天前AI 评分17
去中心化协作持续学习：一种基于多目标最小化的技术
Decentralized collaborative continual learning: A multi-objective minimization-based technique
AI 导读
该研究将去中心化持续学习建模为多目标优化问题，智能体仅做本地计算并与邻居交换信息，通过本地记忆缓冲区存储旧任务样本，将历史信息纳入参数更新以缓解稳定性-可塑性困境。均方误差分析表明，智能体间协作能利用本地数据与记忆缓冲区的多样性，改善跨任务网络平均 MSD。仿真验证了该方法在减少遗忘和提升平均 MSD 上的有效性。
正文 · 原文
Abstract:In this work, we formulate decentralized continual learning within a multi-objective optimization framework. For a given inference task t (corresponding to a common minimizer shared by the cost functions of all agents), agents collecting data in a distributed and streamed manner are only allowed to perform local computations and to exchange information with neighboring agents over the underlying communication graph. As tasks evolve sequentially over time, agents must adapt to newly arriving tasks while retaining knowledge acquired from previously learned ones. This requirement leads to the wellknown stability plasticity dilemma, where stability refers to the ability to retain previous knowledge, while plasticity refers to the ability to learn and adapt to new tasks. To address the stability challenge, agents store subsets of samples from past tasks in local memory buffers. Then, through an appropriate multiobjective formulation, the stored information is incorporated into the learning process so that parameter updates account jointly for the current task and previously learned tasks. The proposed decentralized continual learning approach is analyzed in the mean square error sense under general assumptions on the individual cost functions and gradient noise processes. The analysis reveals that cooperation among agents improves the performance of continual learning. In particular, by exchanging information with neighboring agents, decentralized collaborative learning can exploit the diversity of locally observed data and memory buffers to improve the network average mean-square deviation (MSD) across tasks. Finally, simulations illustrate the theoretical findings and the effectiveness of the method in reducing forgetting and improving the average MSD across tasks.
| Subjects: | Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2610.10882 [cs.MA] |
| (or arXiv:2610.10882v1 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10882 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yara Zgheib [view email]
[v1]
Wed, 7 Oct 2026 20:31:35 UTC (1,903 KB)
来源：arXiv cs.MA · arxiv.org