Tree-of-Ideas: Automated Research Ideation via Cross-Trajectory Reasoning over Scholarly Evolution

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Xun Li, Yiying Yang, Pengtao Li, Xiao Yao, Suyu Liu, Xiaoyang Ye, Ziyu Lu, Yuan Yao, Yangning Li, Yinghui Li, Wenhao Jiang
• Venue: arXiv: cs.AI
• Published: 2026-08-11
• Collected: 2026-08-12
• URL: http://arxiv.org/abs/2608.10740v1
• PDF: https://arxiv.org/pdf/2608.10740v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

The paper introduces Tree-of-Ideas (ToI), a framework to help with research ideation by tracking and analyzing how research problems and solutions evolve across different scholarly trajectories. ToI outperforms existing methods in six AI research topics, scoring higher in novelty and groundedness, and even approaching the score of human-curated paper references.

What the Document Says

• The paper presents Tree-of-Ideas (ToI), a framework that focuses on effective research ideation by examining the evolution of research problems and solutions across scholarly literature.
• EvoTrace is a part of ToI, which reconstructs branching scholarly trajectories from citations, tracking evolving methods, resolved problems, and gaps.
• EvoAgent is another part of ToI that reasons across these trajectories to identify convergent problems and complementary solutions, generating research ideas.
• The performance of ToI was evaluated across six AI research topics.
• ToI achieved the highest score among automatic methods, with a score of 6.27 compared to 5.36 for the strongest baseline on a 10-point scale.
• In terms of specific aspects, ToI scored high in Novelty (6.36) and Groundedness (7.00).
• The score of ToI approaches that of human-curated paper references (6.29), demonstrating the value of cross-path evolutionary reasoning.

Key Concepts Mentioned

• Tree-of-Ideas
• research ideation
• scholarly evolution
• citations
• EvoTrace
• EvoAgent
• AI research topics

Claims Made by the Document

• ToI outperforms existing methods in generating research ideas.
• ToI achieves high scores in Novelty and Groundedness.
• The score of ToI approaches that of human-curated paper references.

Evidence Used

• Across six AI research topics, ToI achieved the highest score among automatic methods (6.27 vs. 5.36 for the strongest baseline on a 10-point scale)
• ToI's score in Novelty was 6.36 and in Groundedness was 7.00.
• The score of ToI approaches that of human-curated paper references (6.29)

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Effective research ideation requires moving beyond a static understanding of prior work to trace how research problems and solutions evolve across the literature. Existing methods either treat papers as unstructured context or model scholarly evolution as isolated citation chains, overlooking interactions among research trajectories. We propose Tree-of-Ideas (ToI), a two-stage framework. EvoTrace reconstructs branching scholarly trajectories from citations, tracking evolving methods, resolved problems, and gaps. EvoAgent then reasons across trajectories to identify convergent problems and complementary solutions, generating grounded research ideas. Across six AI research topics, ToI achieves the highest score among automatic methods (6.27 vs. 5.36 for the strongest baseline on a 10-point scale), with strong Novelty (6.36) and Groundedness (7.00). Also, its score approaches that of human-paper references (6.29), demonstrating the value of cross-path evolutionary reasoning.