Synthesizing scientific literature with retrieval-augmented language models.
Akari AsaiJacqueline HeRulin Shao...Joseph Chee Chang...
Nature·2026·2 citations
TLDROpenScholar is introduced, a specialized retrieval-augmented language model that answers scientific queries by identifying relevant passages from 45 million open-access papers and synthesizing citation-backed responses and improves off-the-shelf LMs by 12%.
How do people cite this paper?
(generated 20 days ago)This work has been recognized as a specialized large language model approach that leverages domain-specific training to improve literature analysis and information retrieval, situating it among efforts to build LLMs tailored for scientific tasks.
Mentions
- Science: Open-source AI program can answer science questions better than humans. Developed by and for academics, OpenScholar aims to improve searches of the ballooning scientific literature. — Jeffrey Brainard
- Ai2 Blog: Scientific literature synthesis with retrieval-augmented language models — Akari Asai
Loading PDF reader...