Exploring Information Seeking Agent Consolidation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915766491676672 |
|---|---|
| author | Yan, Guochen Wu, Jialong Tao, Zhengwei Li, Bo Zhang, Qintong Xu, Jiahao Mi, Haitao Fang, Yuejian Shen, Qingni Zhang, Wentao Wu, Zhonghai |
| author_facet | Yan, Guochen Wu, Jialong Tao, Zhengwei Li, Bo Zhang, Qintong Xu, Jiahao Mi, Haitao Fang, Yuejian Shen, Qingni Zhang, Wentao Wu, Zhonghai |
| contents | Information-seeking agents have emerged as a powerful paradigm for solving knowledge-intensive tasks. Existing information-seeking agents are typically specialized for open web, documents, or local knowledge bases, which constrains scalability and cross-domain generalization. In this work, we investigate how to consolidate heterogeneous information-seeking agents into a single foundation agentic model. We study two complementary consolidation strategies: data-level consolidation, which jointly trains a unified model on a mixture of domain-specific datasets, and parameter-level consolidation, which merges independently trained agent models at the parameter level. Our analysis compares these approaches in terms of performance retention, cross-domain generalization, and interference across information-seeking behaviors. Our results show that data-level consolidation remains a strong and stable baseline, while parameter-level consolidation offers a promising, efficient alternative but suffers from interference and robustness challenges. We further identify key design factors for effective agent consolidation at the parameter level, including fine-grained merging granularity, awareness of task heterogeneity, and principled consensus strategy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_00585 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Exploring Information Seeking Agent Consolidation Yan, Guochen Wu, Jialong Tao, Zhengwei Li, Bo Zhang, Qintong Xu, Jiahao Mi, Haitao Fang, Yuejian Shen, Qingni Zhang, Wentao Wu, Zhonghai Artificial Intelligence Information-seeking agents have emerged as a powerful paradigm for solving knowledge-intensive tasks. Existing information-seeking agents are typically specialized for open web, documents, or local knowledge bases, which constrains scalability and cross-domain generalization. In this work, we investigate how to consolidate heterogeneous information-seeking agents into a single foundation agentic model. We study two complementary consolidation strategies: data-level consolidation, which jointly trains a unified model on a mixture of domain-specific datasets, and parameter-level consolidation, which merges independently trained agent models at the parameter level. Our analysis compares these approaches in terms of performance retention, cross-domain generalization, and interference across information-seeking behaviors. Our results show that data-level consolidation remains a strong and stable baseline, while parameter-level consolidation offers a promising, efficient alternative but suffers from interference and robustness challenges. We further identify key design factors for effective agent consolidation at the parameter level, including fine-grained merging granularity, awareness of task heterogeneity, and principled consensus strategy. |
| title | Exploring Information Seeking Agent Consolidation |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2602.00585 |