Relation Extraction with Instance-Adapted Predicate Descriptions
Yuhang Jiang, MS and Ramakanth Kavuluru, PhD · 2025
Read Relation Extraction with Instance-Adapted Predicate Descriptions by Yuhang Jiang, MS and Ramakanth Kavuluru, PhD (2025) 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/arxiv-2503-17799.txt.
Description
Software that reads text often needs to pull out facts, such as which drug treats which disease. Yuhang Jiang and Ramakanth Kavuluru revisit this task, called relation extraction, in 2025. Large chat models write well, but smaller encoder models still do most of this sorting work, and the pair test a fresh way to train them.
Their design uses two encoders side by side. One reads the text, and the other builds a fresh description of each possible relation using the real names found in that text. Earlier systems used one fixed label for each relation instead. On two medical data sets and two general ones, scores rose by one to two points over the best prior methods.
Catalog record
| CREATOR | Yuhang Jiang, MS and Ramakanth Kavuluru, PhD |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-17799