Learning semantic similarity from sentence pairs using hybrid features centric approach and explainable siamese neural networks
Weihong Zhao, Chunlu Hu · 2026
Read Learning semantic similarity from sentence pairs using hybrid features centric approach and explainable siamese neural networks by Weihong Zhao, Chunlu Hu (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/plos-pone-0345540.txt.
Description
Weihong Zhao and Chunlu Hu tackled a core language problem: judging when two sentences mean the same thing, even when worded very differently. Word-counting methods fail when a short sentence is rephrased with new words. Their model uses twin neural networks that read both sentences, follow the order and context of words, and learn deep patterns of meaning.
Trained on sentence pairs, the model beat word-frequency methods and standard machine learning regressors, with lower errors and steadier results. The twin-network design also allows some explanation of why two sentences scored as similar. Search engines, automatic essay grading, summarizers, and question-answering tools all lean on this kind of similarity judgment, so a more accurate and explainable model has wide use.
Catalog record
| CREATOR | Weihong Zhao, Chunlu Hu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65564 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |