AgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research Agents

2026-05-12

AI Summary

This work addresses the limitation of existing deep research agents, whose tightly coupled exploration and exploitation within a single module hinders efficiency and scalability in open-ended research. The paper proposes a decoupled collaborative architecture that formulates deep research as an adversarial optimization problem between exploration and exploitation for the first time: a critic optimizes query strategies, a generator iteratively refines research outlines, and a report writer produces the final output. A strategy library and meta-optimization mechanism enable automatic strategy discovery and system self-improvement. The authors further introduce GALA, a novel benchmark grounded in real user browsing behavior. Evaluated on three benchmarks, the system matches or surpasses leading closed-source counterparts and is deployed as AutoResearch Your Interest, an end-to-end product offering personalized research recommendations.

Abstract

In this paper, we present AgentDisCo, a novel Disentangled and Collaborative agentic architecture that formulates deep research as an adversarial optimization problem between information exploration and exploitation. Unlike existing approaches that conflate these two processes into a single module, AgentDisCo employs a critic agent to evaluate generated outlines and refine search queries, and a generator agent to retrieve updated results and revise outlines accordingly. The iteratively refined outline is then passed to a downstream report writer that synthesizes a comprehensive research report. The overall workflow supports both handcrafted and automatically discovered design strategies via a meta-optimization harness, in which the generator agent is repurposed as a scoring agent to evaluate critic outputs and generate quality signals. Powerful code-generation agents (e.g., Claude-Code, Codex) systematically explore agent configurations and construct a policy bank, a structured repository of reusable design strategies, enabling the framework to self-refine without extensive human intervention. We evaluate AgentDisCo on three established deep research benchmarks (DeepResearchBench, DeepConsult, DeepResearchGym) using Gemini-2.5-Pro, achieving performance comparable to or surpassing leading closed-source systems. Observing that existing benchmarks inadequately reflect real-world user needs, we introduce GALA (General AI Life Assistants), a benchmark that mines latent research interests from users' historical browsing behavior. We further develop a rendering agent that converts research reports into visually rich poster presentations, and demonstrate an end-to-end product, AutoResearch Your Interest, which delivers personalized deep research recommendations derived from individual browsing histories.

Problem

• disentanglement
• collaboration
• deep research agents
• exploration-exploitation tradeoff
• research benchmarking

Innovation

• disentangled agents
• collaborative optimization
• meta-optimization
• policy bank
• adversarial exploration-exploitation