Literature-Grounded Novelty Assessment of Scientific Ideas
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913914949730304 |
|---|---|
| author | Shahid, Simra Radensky, Marissa Fok, Raymond Siangliulue, Pao Weld, Daniel S. Hope, Tom |
| author_facet | Shahid, Simra Radensky, Marissa Fok, Raymond Siangliulue, Pao Weld, Daniel S. Hope, Tom |
| contents | Automated scientific idea generation systems have made remarkable progress, yet the automatic evaluation of idea novelty remains a critical and underexplored challenge. Manual evaluation of novelty through literature review is labor-intensive, prone to error due to subjectivity, and impractical at scale. To address these issues, we propose the Idea Novelty Checker, an LLM-based retrieval-augmented generation (RAG) framework that leverages a two-stage retrieve-then-rerank approach. The Idea Novelty Checker first collects a broad set of relevant papers using keyword and snippet-based retrieval, then refines this collection through embedding-based filtering followed by facet-based LLM re-ranking. It incorporates expert-labeled examples to guide the system in comparing papers for novelty evaluation and in generating literature-grounded reasoning. Our extensive experiments demonstrate that our novelty checker achieves approximately 13% higher agreement than existing approaches. Ablation studies further showcases the importance of the facet-based re-ranker in identifying the most relevant literature for novelty evaluation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_22026 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Literature-Grounded Novelty Assessment of Scientific Ideas Shahid, Simra Radensky, Marissa Fok, Raymond Siangliulue, Pao Weld, Daniel S. Hope, Tom Information Retrieval Artificial Intelligence I.2; H.3 Automated scientific idea generation systems have made remarkable progress, yet the automatic evaluation of idea novelty remains a critical and underexplored challenge. Manual evaluation of novelty through literature review is labor-intensive, prone to error due to subjectivity, and impractical at scale. To address these issues, we propose the Idea Novelty Checker, an LLM-based retrieval-augmented generation (RAG) framework that leverages a two-stage retrieve-then-rerank approach. The Idea Novelty Checker first collects a broad set of relevant papers using keyword and snippet-based retrieval, then refines this collection through embedding-based filtering followed by facet-based LLM re-ranking. It incorporates expert-labeled examples to guide the system in comparing papers for novelty evaluation and in generating literature-grounded reasoning. Our extensive experiments demonstrate that our novelty checker achieves approximately 13% higher agreement than existing approaches. Ablation studies further showcases the importance of the facet-based re-ranker in identifying the most relevant literature for novelty evaluation. |
| title | Literature-Grounded Novelty Assessment of Scientific Ideas |
| topic | Information Retrieval Artificial Intelligence I.2; H.3 |
| url | https://arxiv.org/abs/2506.22026 |