ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research

arXiv 2026 35.5 benchmark

TLDR

ScholarCatalyst is a benchmark using author-labeled papers to test retrieval of inspiring prior work, showing agentic search underperforms embedding retrieval.

Reasoning

Strengths include a realistic, author-annotated benchmark and clear empirical comparison of retrieval methods. Weaknesses include a narrow CS-paper scope and limited evidence about how retrieval relates to broader scientific discovery.

Read-first score

Read-first score 35.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 28.

Recency 8%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=benchmark,result

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Topical relevance 42%

23.3

Uses existing LLM keyword relevance scores normalized to 0-100. AI scientist,automated scientific discovery,autonomous research agent,automated research,literature review agent,survey generation,automated experimentation,experiment design agent,AI for scientific research,paper writing agent,research automation,scientific discovery agent

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 65.

Keyword Scores

AI for scientific research

literature review agent

autonomous research agent

scientific discovery agent

research automation

automated scientific discovery

automated research

AI scientist

survey generation

automated experimentation

experiment design agent

paper writing agent

Resources

Paper