SentAlign: Accurate and Scalable Sentence Alignment

Steinþór Steingrímsson, Hrafn Loftsson, Andy Way
11/15/2023
cs.CL

No official code found. Best implementation: steinst/SentAlign (38 stars, Apache-2.0, last commit 2026-03-16).

Abstract

We present SentAlign, an accurate sentence alignment tool designed to handle very large parallel document pairs. Given user-defined parameters, the alignment algorithm evaluates all possible alignment paths in fairly large documents of thousands of sentences and uses a divide-and-conquer approach to align documents containing tens of thousands of sentences. The scoring function is based on LaBSE bilingual sentence representations. SentAlign outperforms five other sentence alignment tools when evaluated on two different evaluation sets, German-French and English-Icelandic, and on a downstream machine translation task.

View on arXivView PDF

Code Implementations(1)

steinst/SentAlignCommunity100%
Apache-2.0
389Cython, PythonMar 7, 20237 months agoApache-2.0

Similar Papers

Does mBERT understand Romansh? Evaluating word embeddings using word alignment

85% match1 repo32023

Leveraging Closed-Access Multilingual Embedding for Automatic Sentence Alignment in Low Resource Languages

85% match2023

Adaptative Bilingual Aligning Using Multilingual Sentence Embedding

84% match2024

Multilingual Sentence Transformer as A Multilingual Word Aligner

84% match1 repo242023

WSPAlign: Word Alignment Pre-training via Large-Scale Weakly Supervised Span Prediction

83% match2 repos122023

NAIST-SIC-Aligned: an Aligned English-Japanese Simultaneous Interpretation Corpus

83% match1 repo32023

Cite this paper

@article{steingrmsson2023sentalign,
  title  = {SentAlign: Accurate and Scalable Sentence Alignment},
  author = {Steinþór Steingrímsson and Hrafn Loftsson and Andy Way},
  year   = {2023},
  eprint = {2311.08982},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.08982v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.08982.svg)](https://codeofpaper.com/paper/2311.08982)

Discussion