Juan Pablo Zuluaga-Gomez, Zhaocheng Huang, Xing Niu, Rohit Paturi, Sundararajan Srinivasan, Prashant Mathur, Brian Thompson, Marcello Federico. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Abstract We investigate how humans perform the task of dubbing video content from one language into another, leveraging a novel corpus of 319.57 hours of video from 54 professionally produced titles. This is the first such large-scale study we are aware of. The results challenge a number of assumptions commonly made in both qualitative literature on human dubbing and machine-learning literature on automatic dubbing, arguing for the importance of vocal naturalness and translation quality over commonly emphasized isometric (character length) and lip-sync constraints, and for a more qualified view of the importance of isochronic (timing) constraints. We also find substantial influence of the source-side audio on human dubs through channels other than the words of the translation, pointing to the need for research on ways to preserve speech characteristics, as well as transfer of semantic properties such as emphasis and emotion, in automatic dubbing systems.

Marta Bañón, Pinzhen Chen, Barry Haddow, Kenneth Heafield, Hieu Hoang, Miquel Esplà-Gomis, Mikel L. Forcada, Amir Kamran, Faheem Kirefu, Philipp Koehn, Sergio Ortiz Rojas, Leopoldo Pla Sempere, Gema Ramírez-Sánchez, Elsa Sarrías, Marek Strelec, Brian Thompson, William Waites, Dion Wiggins, Jaume Zaragoza. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.

Many valid translations exist for a given sentence, yet machine translation (MT) is trained with a single reference translation, exacerbating data sparsity in low-resource settings. We introduce Simulated Multiple Reference Training (SMRT), a novel MT training method that approximates the full space of possible translations by sampling a paraphrase of the reference sentence from a paraphraser and training the MT model to predict the paraphraser's distribution over possible tokens. We demonstrate the effectiveness of SMRT in low-resource settings when translating to English, with improvements of 1.2 to 7.0 BLEU. We also find SMRT is complementary to back-translation.

We present a simple document alignment method that incorporates sentence order information in both candidate generation and candidate re-scoring. Our method results in 61% relative reduction in error compared to the best previously published result on the WMT16 document alignment shared task. Our method improves downstream MT performance on web-scraped Sinhala--English documents from ParaCrawl, outperforming the document alignment method used in the most recent ParaCrawl release. It also outperforms a comparable corpora method which uses the same multilingual embeddings, demonstrating that exploiting sentence order is beneficial even if the end goal is sentence-level bitext.

We frame the task of machine translation evaluation as one of scoring machine translation output with a sequence-to-sequence paraphraser, conditioned on a human reference.We propose training the paraphraser as a multilingual NMT system, treating paraphrasing as a zero-shot translation task (e.g., Czech to Czech).This results in the paraphraser's output mode being centered around a copy of the input sequence, which represents the best case scenario where the MT system output matches a human reference.Our method is simple and intuitive, and does not require human judgements for training.Our single model (trained in 39 languages) outperforms or statistically ties with all prior metrics on the WMT 2019 segment-level shared metrics task in all languages (excluding Gujarati where the model had no training data).We also explore using our model for the task of quality estimation as a metric-conditioning on the source instead of the reference-and find that it significantly outperforms every submission to the WMT 2019 shared task on quality estimation in every language pair. Word-level paraphraser log probabilities H(out|in) sBLEU LASERCopy Jason went to school at the University of Madrid . -0.08 -0.26 -0.16 -0.16 -0.12 -0.11 -0.14 -0.10 -0.10 -0.11 -0.10 -0.13 100.0 1.000 Disfluent Jason went school at

Research in machine translation (MT) is developing at a rapid pace. However, most work in the community has focused on languages where large amounts of digital resources are available. In this study, we benchmark state of the art statistical and neural machine translation systems on two African languages which do not have large amounts of resources: Somali and Swahili. These languages are of social importance and serve as test-beds for developing technologies that perform reasonably well despite the low-resource constraint. Our findings suggest that statistical machine translation (SMT) and neural machine translation (NMT) can perform similarly in low-resource scenarios, but neural systems require more careful tuning to match performance. We also investigate how to exploit additional data, such as bilingual text harvested from the web, or user dictionaries; we find that NMT can significantly improve in performance with the use of these additional data. Finally, we survey the landscape of machine translation resources for the languages of Africa and provide some suggestions for promising future research directions.

Brian Thompson, Philipp Koehn. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Brian Thompson, Rebecca Knowles, Xuan Zhang, Huda Khayrallah, Kevin Duh, Philipp Koehn. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

The performance of a spoken language system suffers when non-speech is incorrectly classified as speech. Singing is particularly difficult to discriminate from speech, since both are natural language. However, singing conveys a melody, whereas speech does not; in particular, a singer's fundamental frequency should not deviate significantly from an underlying sequence of notes, while a speaker's fundamental frequency is freer to deviate about a mean value. The present work presents a novel approach to discrimination between singing and speech that exploits the distribution of such deviations. The melody in singing is typically not known a priori, so the distribution cannot be measured directly. Instead, an approximation to its Fourier transform is proposed that allows the unknown melody to be treated as multiplicative noise. This feature vector is shown to be highly discriminative between speech and singing segments when coupled with a simple maximum likelihood classifier, outperforming prior work on real-world data.