Verified, then logged.
Sun 22 Nov 2026 · 00:21 KST / 15:21 UTC한국어 | ENSearch /

Breaking01 / AI

Stored abstracts lift ranking score in paper discovery test

An academic editor viewed from behind compares a manuscript with researcher expertise profiles across two monitors to select potential peer reviewers. AI-generated image

Researchers behind KnowMe Paper Radar reported in a preprint released on 19 Nov 2026 that using stored abstracts raised the tool’s search ranking score to 0.598 (expert-judged retrieval tasks, stored abstracts used). The abstract reports no independent replication results.

According to the abstract, the tool prioritises papers using topics and exclusions set by the reader. It scores papers on how closely they match the topic, how recently they were published and the authors’ research histories.

The researchers measured ranking quality with NDCG@10, which gives more weight to relevant papers near the top of the first 10 results. The score was 0.412 (expert-judged retrieval tasks, stored abstracts unused).

The researchers say the tool displays each part of the score alongside its supporting evidence. In ordinary use, it runs without inference from a large language model.

The abstract does not disclose how many search queries the researchers evaluated.