Language and Voice Lab

Sjá íslenska þýðingu neðar

Last week ACL hosted their 2020 ACL conference. It was supposed to be in Seattle, Washington. But due to COVID-19 it has been moved online, including their satellite events like the ACL 2020 Student Research Workshop (SRW).

This means their proceedings have now been published, including a paper by our very own student, Steinþór Steingrímsson: Effectively Aligning and Filtering Parallel Corpora under Sparse Data Conditions. He wrote about his methods for preparing parallel text for machine translation.

Í síðustu viku hélt ACL árlegu ráðstefnuna sína. Upprunalega átti að halda ráðstefnuna í Seattle en vegna COVID-19 var hún öll færð yfir í netheima, einnig vinnustofur eins og ACL 2020 Student Research Workshop (SRW).

Þetta þýðir að allar innsendar greinar hafa verið gefnar út, þar á meðal grein eftir nemenda okkar Steinþór Steingrímsson. Heiti greinarinnar er Effectively Aligning and Filtering Parallel Corpora under Sparse Data Conditions. Þar skrifar Steinþór um aðferðir til að undirbúa samhliða texta fyrir vélþýðingu.

Abstract

Parallel corpora are key to developing good machine translation systems. However, abundant parallel data are hard to come by, especially for languages with a low number of speakers. When rich morphology exacerbates the data sparsity problem, it is imperative to have accurate alignment and filtering methods that can help make the most of what is available by maximising the number of correctly translated segments in a corpus and minimising noise by removing incorrect translations and segments containing extraneous data. This paper sets out a research plan for improving alignment and filtering methods for parallel texts in low-resource settings. We propose an effective unsupervised alignment method to tackle the alignment problem. Moreover, we propose a strategy to supplement state-of-the-art models with automatically extracted information using basic NLP tools to effectively handle rich morphology.

You can check out the paper on the ACL website: https://www.aclweb.org/anthology/2020.acl-srw.25/

Seattle photo via Good Free Photos