WSPAlign: Word Alignment Pre-training via Large-Scale Weakly Supervised Span Prediction
Qiyu Wu, Masaaki Nagata, Yoshimasa Tsuruoka
Abstract
Most existing word alignment methods rely on manual alignment datasets or parallel corpora, which limits their usefulness. Here, to mitigate the dependence on manual data, we broaden the source of supervision by relaxing the requirement for correct, fully-aligned, and parallel sentences. Specifically, we make noisy, partially aligned, and non-parallel paragraphs. We then use such a large-scale weakly-supervised dataset for word alignment pre-training via span prediction. Extensive experiments with various settings empirically demonstrate that our approach, which is named WSPAlign, is an effective and scalable way to pre-train word aligners without manual data. When fine-tuned on standard benchmarks, WSPAlign has set a new state of the art by improving upon the best supervised baseline by 3.3 6.1 points in F1 and 1.5 6.1 points in AER . Furthermore, WSPAlign also achieves competitive performance compared with the corresponding baselines in few-shot, zero-shot and cross-lingual tests, which demonstrates that WSPAlign is potentially more practical for low-resource languages than existing methods. 1 (1) Data Collection and Annotation (2) Pre-training for word alignment Transformer Encoder Ashikaga Yoshimitsu was the third SEP … ⾜利 義満 は 室町 前期 の ¶ ¶ … Start End Source Paragraph Target Paragraph Alignment training data … Ashikaga Yoshimitsu was the third shogun of the Ashikaga shogunate. ⾜利義満は、室町時代前期の室町幕府第3代征夷⼤将軍 Cross-lingual ⾜利義満は、室町時代前期の… Ashikaga Yoshimitsu was the third shogun… 足利义满在任期间,南北朝获得统一… Ashikaga Yoshimitsu was the third shogun… Ashikaga Yoshimitsu war ein japanischer… Ashikaga unit les cours du Nord et du… Monolingual Ashikaga Yoshimitsu was the third shogun… In 1368 Yoshimitsu was appointed shogun… 足利义满在任期间,南北朝获得统一… 1358年9月25日足利义满出生,当时… Paragraph Pairs Collection Wiki word alignments Common word alignments … … Span Prediction Ashikaga Yoshimitsu was the third shogun… ⾜利義満は、室町時代前期の… Ashikaga Yoshimitsu was the third shogun… ⾜利義満は、室町時代前期の… Wiki Alignment Annotation Ashikaga Yoshimitsu war ein japanischer… Ashikaga Yoshimitsu was the third… Ashikaga Yoshimitsu war ein japanischer… Ashikaga Yoshimitsu was the third… Common Alignment Annotation Co-mentioned entity with Hyperlinks
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext fc20ca77-5545-4b5d-8998-641b617866cdCited by top-tier papers5
- Word Alignment as Preference for Machine TranslationQiyu Wu, Masaaki Nagata, Zhongtao Miao, Yoshimasa TsuruokaEMNLP 2024 · 4 citations
- NeoAMT: Neologism-Aware Agentic Machine Translation with Reinforcement LearningZhongtao Miao, Kaiyan Zhao, Masaaki Nagata, Yoshimasa TsuruokaACL 2026 · 2 citations
- BinaryAlign: Word Alignment as Binary Sequence LabelingGaetan Latouche, Marc-André Carbonneau, Benjamin SwansonACL 2024 · 2 citations
- M²PO: Multi-Perspective Multi-Pair Preference Optimization for Machine TranslationHao Wang, Linlong Xu, Heng Liu, Yangyang Liu et al.ACL 2026
- Towards Fast and Accurate Modeling for Cross-Lingual Label ProjectionThang Le, Huy Huu Nguyen, Anh Tuan Luu, Thamar Solorio et al.ACL 2026
Builds on9
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Robust Speech Recognition via Large-Scale Weak SupervisionAlec Radford, Jong Wook Kim, Tao Xu, Greg Brockman et al.ICML 2023 · 6,966 citations
- Unsupervised Cross-lingual Representation Learning at ScaleAlexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary et al.ACL 2020 · 539 citations
- Rare Words: A Major Problem for Contextualized Embeddings and How to Fix it by Attentive MimickingTimo Schick, Hinrich SchützeAAAI 2020 · 106 citations
- A Supervised Word Alignment Method based on Cross-Language Span Prediction using Multilingual BERTMasaaki Nagata, Katsuki Chousa, Masaaki NishinoEMNLP 2020 · 38 citations
Related papers
- Cross-Align: Modeling Deep Cross-lingual Interactions for Word AlignmentSiyu Lai, Zhen Yang, Fandong Meng, Yufeng Chen et al.EMNLP 2022 · 6 citations
- Improving Pretrained Cross-Lingual Language Models via Self-Labeled Word AlignmentZewen Chi, Li Dong, Bo Zheng, Shaohan Huang et al.ACL 2021
- Mask-Align: Self-Supervised Neural Word AlignmentChi Chen, Maosong Sun, Yang LiuACL 2021
- PreAlign: Boosting Cross-Lingual Transfer by Early Establishment of Multilingual AlignmentJiahuan Li, Shujian Huang, Aarron Ching, Xinyu Dai et al.EMNLP 2024 · 5 citations
- The Reasonableness Behind Unreasonable Translation Capability of Large Language ModelTingchen Fu, Lemao Liu, Deng Cai, Guoping Huang et al.ICLR 2024
