An auditable pipeline for fuzzy full-text screening in systematic reviews: integrating contrastive semantic highlighting and LLM judgment
Arya Rahgozar, Pouria Mortezaagha · 2026
Read An auditable pipeline for fuzzy full-text screening in systematic reviews: integrating contrastive semantic highlighting and LLM judgment by Arya Rahgozar, Pouria Mortezaagha (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13529844.txt.
Description
Arya Rahgozar and Pouria Mortezaagha built an auditable pipeline that speeds full-text screening for systematic reviews. Published in 2026, the method tackles the bottleneck where eligibility clues hide across long, varied papers, tested within a population-health modeling review network. Screening usually demands slow, careful reading of every candidate paper.
Articles are split into overlapping chunks, scored for inclusion versus exclusion similarity with a vagueness margin, and mapped by a fuzzy controller into graded inclusion degrees. A language-model judge then reviews highlighted passages with reasons and confidence, softening scores when evidence runs thin rather than rejecting outright. Pilot recall beat statistical and crisp baselines, with screening time falling from about twenty minutes to under one minute per article. Readers get a traceable shortcut, with full mixed-label validation still ahead.
Catalog record
| CREATOR | Arya Rahgozar, Pouria Mortezaagha |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |