ClaimNov Cuts AI Novelty Check False Rejections to 1.5%
The finding reframes novelty filtering as a recall-versus-precision tradeoff: checking whole proposals punishes incremental builds on prior work, which is how most real papers actually happen. Anyone running an automated research-idea pipeline can adopt the claim-level scoring method directly, keeping proposals that extend one known result. The 96%-versus-1.5% rejection gap gives a concrete benchmark for auditing your own filter.
A University of Illinois paper examines AI scientist pipelines where a novelty check decides which generated research ideas receive compute. The deployed check caught all 380 disguised copies but rejected 96% of ideas that researchers later published. The team built ClaimNov, which splits a proposal into individual claims, matches each claim against claims from prior papers, and gives a judge claim-by-claim evidence. ClaimNov caught 97% of copies while rejecting only 1.5% of the ideas researchers later published.
New University of Illinois paper finds that when AI decides if a research idea is new enough to work on, it often says no to ideas that human scientists later publish.
AI scientist pipelines run a novelty check to decide which generated ideas get compute. These checks judge a whole proposal at once, though most real papers build on some earlier contributions.
They built ClaimNov, which splits a proposal into a few claims, matches each against claims from prior papers, and gives a judge the claim-by-claim evidence.
A deployed check caught all 380 disguised copies but also rejected 96% of ideas researchers later published. ClaimNov caught 97% of copies while rejecting only 1.5% of those ideas.
If your pipeline filters ideas for novelty, score claims separately and keep proposals that build on 1 known result instead of discarding them.
