An Energy-based Model for Word-level AutoCompletion in Computer-aided Translation

Збережено в:
Бібліографічні деталі
Опубліковано в::arXiv.org (Jul 29, 2024), p. n/a
Автор: Cheng, Yang
Інші автори: Huang, Guoping, Yu, Mo, Zhang, Zhirui, Li, Siheng, Yang, Mingming, Shi, Shuming, Yang, Yujiu, Liu, Lemao
Опубліковано:
Cornell University Library, arXiv.org
Предмети:
Онлайн доступ:Citation/Abstract
Full text outside of ProQuest
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!

MARC

LEADER 00000nab a2200000uu 4500
001 3086142275
003 UK-CbPIL
022 |a 2331-8422 
024 7 |a 10.1162/tacl_a_00637  |2 doi 
035 |a 3086142275 
045 0 |b d20240729 
100 1 |a Cheng, Yang 
245 1 |a An Energy-based Model for Word-level AutoCompletion in Computer-aided Translation 
260 |b Cornell University Library, arXiv.org  |c Jul 29, 2024 
513 |a Working Paper 
520 3 |a Word-level AutoCompletion(WLAC) is a rewarding yet challenging task in Computer-aided Translation. Existing work addresses this task through a classification model based on a neural network that maps the hidden vector of the input context into its corresponding label (i.e., the candidate target word is treated as a label). Since the context hidden vector itself does not take the label into account and it is projected to the label through a linear classifier, the model can not sufficiently leverage valuable information from the source sentence as verified in our experiments, which eventually hinders its overall performance. To alleviate this issue, this work proposes an energy-based model for WLAC, which enables the context hidden vector to capture crucial information from the source sentence. Unfortunately, training and inference suffer from efficiency and effectiveness challenges, thereby we employ three simple yet effective strategies to put our model into practice. Experiments on four standard benchmarks demonstrate that our reranking-based approach achieves substantial improvements (about 6.07%) over the previous state-of-the-art model. Further analyses show that each strategy of our approach contributes to the final performance. 
653 |a Labels 
653 |a Machine translation 
653 |a Computer aided mapping 
653 |a Neural networks 
653 |a Context 
653 |a Sentences 
653 |a Effectiveness 
700 1 |a Huang, Guoping 
700 1 |a Yu, Mo 
700 1 |a Zhang, Zhirui 
700 1 |a Li, Siheng 
700 1 |a Yang, Mingming 
700 1 |a Shi, Shuming 
700 1 |a Yang, Yujiu 
700 1 |a Liu, Lemao 
773 0 |t arXiv.org  |g (Jul 29, 2024), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3086142275/abstract/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2407.20083