EpiCon：通过共同演化多模态记忆实现智能体集体学习
EpiCon: Collective Agent Learning through Co-Evolving Multimodal Memory
- MIT-IBM Computing Research Lab（MIT-IBM计算研究实验室）
- University of Rochester（罗切斯特大学）
- Massachusetts Institute of Technology（麻省理工学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
EpiCon提出共享多模态记忆框架，通过记忆控制器和树状自组织器实现智能体集体学习，在十一个基准上将宏平均分数提升1.7至4.9分，并显著降低记忆操作时间。
AI中文摘要：
智能体可以从过去的执行中学习，但让不同智能体复用并基于彼此的经验进行构建仍然具有挑战性。我们提出了EpiCon，一个共享的多模态记忆框架，用于智能体集体学习，无需更新宿主模型参数。EpiCon通过两个独立训练的2B模型（一个记忆控制器和一个树状自组织器）将问题级别的记忆演化与持久经验库连接起来。控制器在多次尝试中联合优化文本指导和视觉证据，并选择性纳入视觉记忆。自组织器分层整合经验教训，并为新问题检索经验和规则。我们在涵盖四个多模态任务域的十一个基准上评估了EpiCon，使用了两个测试框架和多个主干模型。一个冻结的经验库即使在单次求解尝试中也能提升其他系统的性能。第二个测试框架将原始系统的宏平均分数在十一个基准上提高了2.6分。在四种宿主配置中，与无记忆相比，EpiCon将宏平均分数提高了1.7至4.9分，并将记忆操作时间相对于主干规模记忆模型减少了67%至74%。
英文摘要：
Agents can learn from past executions, but enabling different agents to reuse and build on one another's experience remains challenging. We introduce EpiCon, a shared multimodal memory framework for agent collective learning without updating host model parameters. EpiCon links question-level memory evolution to a persistent experience bank through two independently trained 2B models: a memory controller and a tree self-organizer. The controller jointly refines textual guidance and visual evidence across attempts and selectively includes visual memory. The self-organizer consolidates lessons hierarchically and retrieves experience and rules for new problems. We evaluate EpiCon on eleven benchmarks spanning four multimodal task domains, using two harnesses and multiple backbones. A frozen bank improves other systems even with a single solving attempt. A second harness raises the original system's macro-average score by 2.6 points across eleven benchmarks. Across four host configurations, EpiCon improves macro-average scores by 1.7 to 4.9 points over No Memory and reduces memory-operation time by 67\% to 74\% relative to backbone-sized memory models.
