![[2602.23258] AgentDropoutV2: Optimizing Information Flow in Multi-Agent Systems via Test-Time Rectify-or-Reject Pruning](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.23258] AgentDropoutV2: Optimizing Information Flow in Multi-Agent Systems via Test-Time Rectify-or-Reject Pruning
Summary
AgentDropoutV2 introduces a novel pruning framework to enhance information flow in Multi-Agent Systems by dynamically correcting errors during test time, improving accuracy and adaptability.
Why It Matters
This research addresses the critical issue of error propagation in Multi-Agent Systems, which can hinder performance in complex tasks. By proposing a solution that optimizes information flow without the need for retraining, it enhances the deployability and adaptability of these systems, making them more effective in real-world applications.
Key Takeaways
- AgentDropoutV2 improves Multi-Agent Systems by correcting errors dynamically at test time.
- The framework uses a retrieval-augmented rectifier to enhance task performance by an average of 6.3 percentage points.
- It prevents error propagation by pruning irreparable outputs, maintaining system integrity.
- The system adapts rectification efforts based on task difficulty and context-aware indicators.
- The authors provide code and datasets for further research and application.
Computer Science > Artificial Intelligence arXiv:2602.23258 (cs) [Submitted on 26 Feb 2026] Title:AgentDropoutV2: Optimizing Information Flow in Multi-Agent Systems via Test-Time Rectify-or-Reject Pruning Authors:Yutong Wang, Siyuan Xiong, Xuebo Liu, Wenkang Zhou, Liang Ding, Miao Zhang, Min Zhang View a PDF of the paper titled AgentDropoutV2: Optimizing Information Flow in Multi-Agent Systems via Test-Time Rectify-or-Reject Pruning, by Yutong Wang and 6 other authors View PDF HTML (experimental) Abstract:While Multi-Agent Systems (MAS) excel in complex reasoning, they suffer from the cascading impact of erroneous information generated by individual participants. Current solutions often resort to rigid structural engineering or expensive fine-tuning, limiting their deployability and adaptability. We propose AgentDropoutV2, a test-time rectify-or-reject pruning framework designed to dynamically optimize MAS information flow without retraining. Our approach acts as an active firewall, intercepting agent outputs and employing a retrieval-augmented rectifier to iteratively correct errors based on a failure-driven indicator pool. This mechanism allows for the precise identification of potential errors using distilled failure patterns as prior knowledge. Irreparable outputs are subsequently pruned to prevent error propagation, while a fallback strategy preserves system integrity. Empirical results on extensive math benchmarks show that AgentDropoutV2 significantly boosts the MAS...
