Philipp Koehn · page 2
Papers
- Re-evaluating the Role of Bleu in Machine Translation Research (2006 · 296 citations)
The authors question whether the Bleu metric really reflects human judgments of machine translation quality well. - CCAligned: A Massive Collection of Cross-Lingual Web-Document Pairs (2020 · 231 citations)
Finding web pages that translate one another is valuable for machine translation and other cross-lingual NLP work. - Findings of the 2012 Workshop on Statistical Machine Translation (2012 · 204 citations)
This report covers the shared tasks of WMT12, the seventh Workshop on Statistical Machine Translation, held at NAACL 2012. - On the Impact of Various Types of Noise on Neural Machine Translation (2018 · 188 citations)
The paper asks which kinds of flaws in parallel training data damage translation quality, and whether neural and statistical systems react differently. - Overcoming Catastrophic Forgetting During Domain Adaptation of Neural Machine Translation (2019 · 185 citations)
When a general-domain neural machine translation model is fine-tuned on a small in-domain corpus (continued training), it improves on the target domain but translates generic text much worse. - Findings of the WMT 2018 Shared Task on Parallel Corpus Filtering (2018 · 181 citations)
The WMT 2018 parallel corpus filtering task asks how to pick out usable sentence pairs from extremely noisy web-crawled bitext. - Dirt Cheap Web-Scale Parallel Text from the Common Crawl (2013 · 166 citations)
Statistical machine translation depends on parallel text. Curated corpora cover only a few language pairs, and most of them are government or news text. Large web crawls could supply more, but only big companies have been able to crawl and process the web at that scale. The authors avoid crawling altogether by mining… - Aiding Pronoun Translation with Co-Reference Resolution (2010 · 165 citations)
Statistical MT systems translate each sentence independently, so they cannot choose the correct gender for English it and they when the antecedent sits in an earlier sentence. - Convergence of Translation Memory and Statistical Machine Translation (2010 · 153 citations)
Translation memories (TM) and statistical machine translation (SMT) grew up in separate communities, even though translation agencies increasingly want to use both. - Statistical Power and Translationese in Machine Translation Evaluation (2020 · 150 citations)
This paper asks how reliable machine translation (MT) evaluation is when test sets mix two kinds of source text. - Vecalign: Improved Sentence Alignment in Linear Time and Space (2019 · 144 citations)
Sentence alignment pairs up minimal groups of sentences in parallel documents that translate each other. Older aligners are slow on long documents because their search grows quadratically, and stronger systems such as Bleualign rely on an MT system. Thompson and Koehn present Vecalign, an aligner that needs only… - Findings of the 2011 Workshop on Statistical Machine Translation (2011 · 136 citations)
This report documents the outcomes of the WMT11 shared tasks: translation between English and Czech, French, German and Spanish, combining outputs of several MT systems, and evaluation metrics judged against human preferences. - TICO-19: the Translation Initiative for COvid-19 (2020 · 133 citations)
The TICO-19 project addresses a communication gap during the COVID-19 pandemic. Authoritative health guidance existed mostly in major languages, so speakers of many low-resource languages were left exposed to misinformation. A coalition of academic groups, technology companies, language service providers and… - Findings of the WMT 2019 Shared Task on Parallel Corpus Filtering for Low-Resource Conditions (2019 · 132 citations)
Methods that filter noisy web-crawled bitext work well for high-resource languages, but it was unclear how they behave when very little clean data exists to train the scoring models. - Low-Resource Corpus Filtering Using Multilingual Sentence Embeddings (2019 · 110 citations)
This paper presents a submission to the WMT19 shared task on filtering noisy web-crawled parallel data for low-resource language pairs. - Findings of the WMT 2020 Shared Task on Parallel Corpus Filtering and Alignment (2020 · 101 citations)
Web-crawled parallel text is plentiful but very noisy, and that noise hurts neural MT. This matters most for low-resource pairs, where little clean data exists. The third WMT shared task on this problem targeted Pashto–English and Khmer–English. It also added a new subtask: aligning sentences within crawled document… - The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contexts (2024 · 88 citations)
This paper asks whether the safety behaviour of large language models holds up when the same harmful request is posed in languages with little training data. - SimulMT to SimulST: Adapting Simultaneous Text Translation to End-to-End Simultaneous Speech Translation (2020 · 81 citations)
Methods for simultaneous text translation (SimulMT) decide token by token when to read input and when to emit output. - Facebook AI's WMT21 News Translation Task Submission (2021 · 42 citations)
The paper asks whether one multilingual system can beat strong, carefully built bilingual systems on competitive news translation. - A Survey of Qualitative Error Analysis for Neural Machine Translation Systems (2020)
The paper is a qualitative review of errors made by neural MT systems in production, across 33 languages from 10 families.
This page in Markdown → https://md.research.elamur.ai/free/en/authors/philipp-koehn/page/2/