AMATA: Adaptive Multi-Agent Trajectory Alignment for Knowledge-Intensive Question Answering
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910228877934592 |
|---|---|
| author | Zhang, Taolin Li, Dongyang Chen, Chen Chen, Qizhou Wan, Jiuheng He, Xiaofeng Wang, Chengyu Hong, Richang |
| author_facet | Zhang, Taolin Li, Dongyang Chen, Chen Chen, Qizhou Wan, Jiuheng He, Xiaofeng Wang, Chengyu Hong, Richang |
| contents | Despite substantial advances in large language models (LLMs), generating factually consistent responses for knowledge-intensive question answering remains challenging. These difficulties are primarily due to hallucinations and the limitations of LLMs in bridging long-tail knowledge gaps. To address this, we propose AMATA, an Adaptive Multi-Agent Trajectory Alignment framework that dynamically integrates external knowledge to improve response interpretability and factual grounding. Our architecture leverages six specialized agents that collaboratively perform structured actions for complex question reasoning. We formalize multi-agent collaboration with external tools as a trajectory preference alignment problem, incorporating question-aware agent customization and inter-agent preference harmonization. AMATA introduces two principal innovations: (1) Intra-Trajectory Preference Learning, which learns objective-oriented preferences to prioritize critical agents, and (2) Inter-Agent Dependency Learning, which captures cross-agent tool dependencies through a novel dependency-aware direct preference optimization technique. Empirical results show that AMATA consistently outperforms baseline approaches, knowledge-augmented frameworks, and LLM-based trajectory systems on five established knowledge-intensive QA benchmarks. Further analysis demonstrates the efficiency of our method in reducing token consumption. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_17352 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AMATA: Adaptive Multi-Agent Trajectory Alignment for Knowledge-Intensive Question Answering Zhang, Taolin Li, Dongyang Chen, Chen Chen, Qizhou Wan, Jiuheng He, Xiaofeng Wang, Chengyu Hong, Richang Computation and Language Despite substantial advances in large language models (LLMs), generating factually consistent responses for knowledge-intensive question answering remains challenging. These difficulties are primarily due to hallucinations and the limitations of LLMs in bridging long-tail knowledge gaps. To address this, we propose AMATA, an Adaptive Multi-Agent Trajectory Alignment framework that dynamically integrates external knowledge to improve response interpretability and factual grounding. Our architecture leverages six specialized agents that collaboratively perform structured actions for complex question reasoning. We formalize multi-agent collaboration with external tools as a trajectory preference alignment problem, incorporating question-aware agent customization and inter-agent preference harmonization. AMATA introduces two principal innovations: (1) Intra-Trajectory Preference Learning, which learns objective-oriented preferences to prioritize critical agents, and (2) Inter-Agent Dependency Learning, which captures cross-agent tool dependencies through a novel dependency-aware direct preference optimization technique. Empirical results show that AMATA consistently outperforms baseline approaches, knowledge-augmented frameworks, and LLM-based trajectory systems on five established knowledge-intensive QA benchmarks. Further analysis demonstrates the efficiency of our method in reducing token consumption. |
| title | AMATA: Adaptive Multi-Agent Trajectory Alignment for Knowledge-Intensive Question Answering |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.17352 |