Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Tale of Two Graphs: Separating Knowledge Exploration from Outline Structure for Open-Ended Deep Research

Zhuofan Shi, Ming Ma, Zekun Yao, Fangkai Yang, Jue Zhang, Dongge Han, Victor Rühle, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang

TL;DR

DualGraph memory is presented, an architecture that separates what the agent knows from how it writes, enabling more efficient and comprehensive iterative knowledge-driven exploration and refinement in OEDR agents.

Abstract

Open-Ended Deep Research (OEDR) pushes LLM agents beyond short-form QA toward long-horizon workflows that iteratively search, connect, and synthesize evidence into structured reports. However, existing OEDR agents largely follow either linear ``search-then-generate'' accumulation or outline-centric planning. The former suffers from lost-in-the-middle failures as evidence grows, while the latter relies on the LLM to implicitly infer knowledge gaps from the outline alone, providing weak supervision for identifying missing relations and triggering targeted exploration. We present DualGraph memory, an architecture that separates what the agent knows from how it writes. DualGraph maintains two co-evolving graphs: an Outline Graph (OG), and a Knowledge Graph (KG), a semantic memory that stores fine-grained knowledge units, including core entities, concepts, and their relations. By analyzing the KG topology together with structural signals from the OG, DualGraph generates targeted search queries, enabling more efficient and comprehensive iterative knowledge-driven exploration and refinement. Across DeepResearch Bench, DeepResearchGym, and DeepConsult, DualGraph consistently outperforms state-of-the-art baselines in report depth, breadth, and factual grounding; for example, it reaches a 53.08 RACE score on DeepResearch Bench with GPT-5. Moreover, ablation studies confirm the central role of the dual-graph design.

A Tale of Two Graphs: Separating Knowledge Exploration from Outline Structure for Open-Ended Deep Research

TL;DR

DualGraph memory is presented, an architecture that separates what the agent knows from how it writes, enabling more efficient and comprehensive iterative knowledge-driven exploration and refinement in OEDR agents.

Abstract

Open-Ended Deep Research (OEDR) pushes LLM agents beyond short-form QA toward long-horizon workflows that iteratively search, connect, and synthesize evidence into structured reports. However, existing OEDR agents largely follow either linear ``search-then-generate'' accumulation or outline-centric planning. The former suffers from lost-in-the-middle failures as evidence grows, while the latter relies on the LLM to implicitly infer knowledge gaps from the outline alone, providing weak supervision for identifying missing relations and triggering targeted exploration. We present DualGraph memory, an architecture that separates what the agent knows from how it writes. DualGraph maintains two co-evolving graphs: an Outline Graph (OG), and a Knowledge Graph (KG), a semantic memory that stores fine-grained knowledge units, including core entities, concepts, and their relations. By analyzing the KG topology together with structural signals from the OG, DualGraph generates targeted search queries, enabling more efficient and comprehensive iterative knowledge-driven exploration and refinement. Across DeepResearch Bench, DeepResearchGym, and DeepConsult, DualGraph consistently outperforms state-of-the-art baselines in report depth, breadth, and factual grounding; for example, it reaches a 53.08 RACE score on DeepResearch Bench with GPT-5. Moreover, ablation studies confirm the central role of the dual-graph design.
Paper Structure (47 sections, 9 equations, 4 figures, 4 tables, 2 algorithms)

This paper contains 47 sections, 9 equations, 4 figures, 4 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Works
  3. Open-Ended Deep Research Agent
  4. Graph-Based Memory for Deep Research
  5. Methodology
  6. Preliminary
  7. Overview of DualGraph
  8. Co-Evolution of Outline and Knowledge Graph
  9. Report Generation
  10. Experiments
  11. Experimental Setup
  12. Benchmarks.
  13. Metrics.
  14. Baselines.
  15. Implementation Details.
  16. ...and 32 more sections

Figures (4)

  • Figure 1: Motivating example contrasting DualGraph (left) and outline-centric (right) approaches. Starting from the shared outline graph, the outline-centric baseline continues to elaborate Duan's philosophy generally, fails to explore whether Duan emphasizes opportunity cost, resulting in an incomplete comparison. In contrast, DualGraph exposes the missing Duan–opportunity cost relation in the Knowledge Graph and triggers a targeted in-depth query for a more comprehensive analysis.
  • Figure 2: Overview of our DualGraph framework. shows the OG initialization, - shows the co-evolving iteration of KG and OG, and shows the report finalization after the iteration termination.
  • Figure 3: LLM-as-a-judge scores of intermediate outlines generated by DualGraph and DualGraph w/o KG across optimization iterations on DeepResearch Bench.
  • Figure 4: LLM-as-a-judge comparison of intermediate outlines generated by DualGraph and DualGraph w/o KG across optimization iterations on DeepResearch Bench.