基于合作驱动优化动态的多智能体学习

AI 摘要

该论文提出一种多智能体协作训练机制，让多个参数量较少的「小」神经网络在训练过程中共享预测结果，并将这些信息融入损失函数以直接影响权重更新。作者测试了投票模型、多数模型和基于置信度的加权平均模型等协作策略，在多个标准基准上进行了数值比较。结果显示，多个小智能体在给定分类任务上可以超越单个大模型，同时显著减少需训练的参数数量，从而降低计算资源消耗。

正文节选

Multi-Agent Learning with Cooperation-Driven Optimization Dynamics Abstract Multilayer Artificial Neural Networks trained via backpropagation are the basic blocks of many, more complex, classification algorithms. Their strength lies in the possibility of realizing, with arbitrary precision, any function. This result comes at the cost of the large number of involved parameters to be optimized. In this work, we propose a mechanism for cooperation, i.e., information exchange among several artificia

发布时间：2026-09-16 12:00

抓取时间：2026-09-16 12:25

来源机构：arXiv