aiqbench analysis
This research introduces DUOTRACE, a novel framework designed to enhance the reliability of LLM-based multi-agent systems by implementing a "detect-before-attribute" paradigm for failure analysis. By combining dual-view semantic representations with Tree-LSTM encoders, the method effectively addresses long-context degradation and improves attribution accuracy across various agent trajectories. The study highlights critical advancements in maintaining system integrity as autonomous agents scale to handle increasingly complex reasoning tasks.
Detect Before You Attribute: Cascade Failure Attribution for Multi-Agent Systems
arXiv:2608.29646v1 Announce Type: new Abstract: Large language model (LLM)-based agents have shown strong potential in solving complex tasks through multi-step reasoning, yet they remain vulnerable to execution failures. Accurate failure attribution is therefore critical for improving agent reliability. Existing topology- and spectrum-based methods exploit trajectory structures but often overlook fine-grained semantics, while LLM-based attribution methods capture semantic cues but suffer from long-context degradation over lengthy trajectories. To address these challenges, we propose DUOTRACE, a plug-and-play detection filter for LLM-based failure attribution. DUOTRACE follows a detect-before-attribute paradigm: it first detects anomalous executions and then supplies focused trajectory evidence to downstream LLM-based attribution methods. For effective VAE-based anomaly detection on agent trajectories, DUOTRACE integrates dual-view semantic-structural node representations, a Tree-LSTM-based trajectory encoder, and prefix-chain- and LLM-based data augmentation to handle heterogeneous nodes, hierarchical execution structures, and limited failure data. Experiments with six LLM-based attribution baselines show that DUOTRACE improves agent-level and step-level attribution accuracy by 8.7% and 7.0%, respectively.
Abstract:Large language model (LLM)-based agents have shown strong potential in solving complex tasks through multi-step reasoning, yet they remain vulnerable to execution failures. Accurate failure attribution is therefore critical for improving agent reliability. Existing topology- and spectrum-based methods exploit trajectory structures but often overlook fine-grained semantics, while LLM-based attribution methods capture semantic cues but suffer from long-context degradation over lengthy trajectories. To address these challenges, we propose DUOTRACE, a plug-and-play detection filter for LLM-based failure attribution. DUOTRACE follows a detect-before-attribute paradigm: it first detects anomalous executions and then supplies focused trajectory evidence to downstream LLM-based attribution methods. For effective VAE-based anomaly detection on agent trajectories, DUOTRACE integrates dual-view semantic-structural node representations, a Tree-LSTM-based trajectory encoder, and prefix-chain- and LLM-based data augmentation to handle heterogeneous nodes, hierarchical execution structures, and limited failure data. Experiments with six LLM-based attribution baselines show that DUOTRACE improves agent-level and step-level attribution accuracy by 8.7% and 7.0%, respectively.
Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
arXiv:2608.29646 [cs.AI]
(or arXiv:2608.29646v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.29646 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 249 words
Related coverage
Agent Evaluation Metric for multi-turn conversations
AWS · Sep 10, 2026
OrchestraBench: Evaluating Multi-Agent Orchestration Failure Modes, Recovery, and Decomposition Quality
arXiv · Aug 7, 2026
Article: The Platform Engineering Playbook for Production LLMs
InfoQ · Oct 5, 2026
TRAJDEBUG: Tracing Error Lifecycle to Identify Critical Failures in Long-Horizon Agent Trajectories
arXiv · Aug 7, 2026