A parallel dual-stream state-space module for reliable and efficient biomedical relation extraction
Yaxun Jia, Zhu Yuan, Lian Zhu, Bing Han, Li Ren, Zuo-lin Xiang · 2026
Read A parallel dual-stream state-space module for reliable and efficient biomedical relation extraction by Yaxun Jia, Zhu Yuan, Lian Zhu, Bing Han, Li Ren, Zuo-lin Xiang (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/PMC13423190.txt.
Description
Jia and colleagues present in 2026 DuSSM, a computer system that extracts drug-drug interactions from biomedical text, a backbone task for drug safety monitoring worldwide. The challenge is that local word cues can distort the long-range logic of a sentence.
The design splits the work into two parallel streams. A convolutional stream scans for local trigger patterns, while a state-space stream tracks the sentence's evolving meaning in a single efficient pass. Tested on four biomedical benchmarks, the system reached an F1 score in the low eighty percent range on the main drug interaction benchmark. An F1 score combines precision and recall, and the system scored notably well on precision for sentence pairs. Precision on cases with no interaction was especially high, which matters because false alerts tire the pharmacists who must review them.
Catalog record
| CREATOR | Yaxun Jia, Zhu Yuan, Lian Zhu, Bing Han, Li Ren, Zuo-lin Xiang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |