English
Related papers

Related papers: Unsupervised Parallel Corpus Mining on Web Data

200 papers

We show how to derive state-of-the-art unsupervised neural machine translation systems from generatively pre-trained language models. Our method consists of three steps: few-shot amplification, distillation, and backtranslation. We first…

Pre-training has proven to be effective in unsupervised machine translation due to its ability to model deep context information in cross-lingual scenarios. However, the cross-lingual information obtained from shared BPE spaces is…

Computation and Language · Computer Science 2019-09-04 Shuo Ren , Yu Wu , Shujie Liu , Ming Zhou , Shuai Ma

We present an approach based on multilingual sentence embeddings to automatically extract parallel sentences from the content of Wikipedia articles in 85 languages, including several dialects or low-resource languages. We do not limit the…

Computation and Language · Computer Science 2019-07-17 Holger Schwenk , Vishrav Chaudhary , Shuo Sun , Hongyu Gong , Francisco Guzmán

A crowdsourcing translation approach is an effective tool for globalization of site content, but it is also an important source of parallel linguistic data. For the given site, processed with a crowdsourcing system, a sentence-aligned…

Computation and Language · Computer Science 2014-09-22 Alexander Kalinin , George Savchenko

The effectiveness of a statistical machine translation system (SMT) is very dependent upon the amount of parallel corpus used in the training phase. For low-resource language pairs there are not enough parallel corpora to build an accurate…

Computation and Language · Computer Science 2017-01-31 Ebrahim Ansari , M. H. Sadreddini , Mostafa Sheikhalishahi , Richard Wallace , Fatemeh Alimardani

Illegal marketplaces have increasingly shifted to concealed parts of the internet, including the deep and dark web, as well as platforms such as Telegram, Reddit, and Pastebin. These channels enable the anonymous trade of illicit goods…

Despite impressive progress in high-resource settings, Neural Machine Translation (NMT) still struggles in low-resource and out-of-domain scenarios, often failing to match the quality of phrase-based translation. We propose a novel…

Computation and Language · Computer Science 2018-05-31 Xing Niu , Michael Denkowski , Marine Carpuat

Without real bilingual corpus available, unsupervised Neural Machine Translation (NMT) typically requires pseudo parallel data generated with the back-translation method for the model training. However, due to weak supervision, the pseudo…

Computation and Language · Computer Science 2019-01-15 Shuo Ren , Zhirui Zhang , Shujie Liu , Ming Zhou , Shuai Ma

Expert-layman text style transfer technologies have the potential to improve communication between members of scientific communities and the general public. High-quality information produced by experts is often filled with difficult jargon…

Computation and Language · Computer Science 2021-12-21 Wenda Xu , Michael Saxon , Misha Sra , William Yang Wang

In this paper, we introduce a novel approach to generate synthetic data for training Neural Machine Translation systems. The proposed approach transforms a given parallel corpus between a written language and a target language to a parallel…

Computation and Language · Computer Science 2017-11-30 Hany Hassan , Mostafa Elaraby , Ahmed Tawfik

In this paper, we investigate the problem of training neural machine translation (NMT) systems with a dataset of more than 40 billion bilingual sentence pairs, which is larger than the largest dataset to date by orders of magnitude.…

Computation and Language · Computer Science 2019-10-08 Yuxian Meng , Xiangyuan Ren , Zijun Sun , Xiaoya Li , Arianna Yuan , Fei Wu , Jiwei Li

Quality Estimation (QE) is the task of evaluating the quality of a translation when reference translation is not available. The goal of QE aligns with the task of corpus filtering, where we assign the quality score to the sentence pairs…

Computation and Language · Computer Science 2023-06-07 Akshay Batheja , Pushpak Bhattacharyya

Translation is important for cross-language communication, and many efforts have been made to improve its accuracy. However, less investment is conducted in aligning translations with human preferences, such as translation tones or styles.…

Computation and Language · Computer Science 2024-10-16 Shuqiao Sun , Yutong Yao , Peiwen Wu , Feijun Jiang , Kaifu Zhang

Paraphrasing exemplifies the ability to abstract semantic content from surface forms. Recent work on automatic paraphrasing is dominated by methods leveraging Machine Translation (MT) as an intermediate step. This contrasts with humans, who…

Machine Learning · Computer Science 2019-05-31 Aurko Roy , David Grangier

We present the HPLT (High Performance Language Technologies) language resources, a new massive multilingual dataset including both monolingual and bilingual corpora extracted from CommonCrawl and previously unused web crawls from the…

Analogy-making is central to human cognition, allowing us to adapt to novel situations -- an ability that current AI systems still lack. Most analogy datasets today focus on simple analogies (e.g., word analogies); datasets including…

Computation and Language · Computer Science 2024-05-15 Oren Sultan , Yonatan Bitton , Ron Yosef , Dafna Shahaf

Unsupervised neural machine translation (UNMT) is beneficial especially for low resource languages such as those from the Dravidian family. However, UNMT systems tend to fail in realistic scenarios involving actual low resource languages.…

Computation and Language · Computer Science 2021-03-31 Sai Koneru , Danni Liu , Jan Niehues

Neural networks are capable of translating between languages -- in some cases even between two languages where there is little or no access to parallel translations, in what is known as Unsupervised Machine Translation (UMT). Given this…

Computation and Language · Computer Science 2023-11-07 Shafi Goldwasser , David F. Gruber , Adam Tauman Kalai , Orr Paradise

Advancements in unsupervised machine translation have enabled the development of machine translation systems that can translate between languages for which there is not an abundance of parallel data available. We explored unsupervised…

Computation and Language · Computer Science 2023-01-11 Megan Dare , Valentina Fajardo Diaz , Averie Ho Zoen So , Yifan Wang , Shibingfeng Zhang

Word alignments are useful for tasks like statistical and neural machine translation (NMT) and cross-lingual annotation projection. Statistical word aligners perform well, as do methods that extract alignments jointly with translations in…

Computation and Language · Computer Science 2021-04-19 Masoud Jalili Sabet , Philipp Dufter , François Yvon , Hinrich Schütze