\textsc{MasFACT}: Continual Multi-Agent Topology Learning via Geometry-Aware Posterior Transfer

2026-05-17

AI Summary

This work addresses the problem of “topological forgetting” in multi-agent continual learning, where communication structures required for previously learned tasks are overwritten during new task acquisition. The study is the first to identify and tackle this issue by introducing a geometry-aware posterior transfer mechanism. Specifically, it leverages Fused Gromov-Wasserstein optimal transport to align agent interaction spaces across tasks, thereby constructing transferable topological prior representations. A conservative posterior adaptation strategy, theoretically grounded in PAC-Bayes bounds, is then employed to balance topological stability with task adaptability. The proposed method significantly improves average accuracy across diverse continual learning settings, effectively mitigates topological forgetting, and seamlessly integrates with various topology generators.

Abstract

Multi-agent systems (MAS) powered by large language models (LLMs) have emerged as a powerful paradigm for complex problem solving, where performance critically depends on the underlying inter-agent communication topology. However, existing topology generation methods mainly optimize for isolated tasks, while real-world deployments involve streams of evolving tasks, requiring previously effective collaboration patterns to be retained and reused rather than rediscovered or overwritten. We identify a previously underexplored failure mode, \emph{topology forgetting}, in which adapting to new tasks shifts the topology generator away from communication structures required by earlier tasks. This issue stems from cross-task misalignment in both agent-level functional semantics and relational communication structures. To address this challenge, we propose \textbf{\textsc{MasFACT}}, a geometry-aware posterior transfer framework that preserves and reuses historical collaboration knowledge as transferable topology priors. We transfer these priors across task-specific agent spaces through Fused Gromov-Wasserstein optimal transport and perform PAC-Bayes-guided conservative posterior adaptation to balance task-specific plasticity with structural stability. Experiments across class-, domain-, and task-level continual settings demonstrate that \textsc{MasFACT} consistently improves average accuracy while reducing topology forgetting compared to strong topology generation and replay-based baselines, and can be seamlessly integrated with different MAS topology generators.

Problem

• continual learning
• multi-agent systems
• topology forgetting
• communication topology
• task streams

Innovation

• topology forgetting
• geometry-aware posterior transfer
• Fused Gromov-Wasserstein
• PAC-Bayes
• continual multi-agent learning