1 code implementation • IWSLT (ACL) 2022 • Ioannis Tsiamas, Gerard I. Gállego, Carlos Escolano, José Fonollosa, Marta R. Costa-jussà
We further investigate the suitability of different speech encoders (wav2vec 2. 0, HuBERT) for our models and the impact of knowledge distillation from the Machine Translation model that we use for the decoder (mBART).
no code implementations • WMT (EMNLP) 2020 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa
In this article, we describe the TALP-UPC participation in the WMT20 news translation shared task for Tamil-English.
no code implementations • WMT (EMNLP) 2021 • Carlos Escolano, Ioannis Tsiamas, Christine Basta, Javier Ferrando, Marta R. Costa-Jussa, José A. R. Fonollosa
We fine-tune mBART50 using the filtered data, and additionally, we train a Transformer model on the same data from scratch.
no code implementations • ICON 2021 • Carlos Escolano, Graciela Ojeda, Christine Basta, Marta R. Costa-Jussa
Machine Translation is highly impacted by social biases present in data sets, indicating that it reflects and amplifies stereotypes.
1 code implementation • 29 Sep 2023 • Casimiro Pio Carrino, Carlos Escolano, José A. R. Fonollosa
Our approach seeks to enhance cross-lingual QA transfer using a high-performing multilingual model trained on a large-scale dataset, complemented by a few thousand aligned QA examples across languages.
1 code implementation • 19 May 2023 • Javier García Gilabert, Carlos Escolano, Marta R. Costa-jussà
Our proposed method, ReSeTOX (REdo SEarch if TOXic), addresses the issue of Neural Machine Translation (NMT) generating translation outputs that contain toxic words not present in the input.
no code implementations • 6 Oct 2022 • Marta R. Costa-jussà, Eric Smith, Christophe Ropers, Daniel Licht, Jean Maillard, Javier Ferrando, Carlos Escolano
We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages.
1 code implementation • 23 May 2022 • Javier Ferrando, Gerard I. Gállego, Belen Alastruey, Carlos Escolano, Marta R. Costa-jussà
In Neural Machine Translation (NMT), each token prediction is conditioned on the source sentence and the target prefix (what has been previously translated at a decoding step).
1 code implementation • 12 Feb 2022 • Oriol Domingo, Marta R. Costa-jussà, Carlos Escolano
The proposed solution, a T5 architecture, is trained in a multi-task semi-supervised environment, with our collected non-parallel data, following a cycle training regime.
1 code implementation • ACL (IWSLT) 2021 • Gerard I. Gállego, Ioannis Tsiamas, Carlos Escolano, José A. R. Fonollosa, Marta R. Costa-jussà
Our submission also uses a custom segmentation algorithm that employs pre-trained Wav2Vec 2. 0 for identifying periods of untranscribable text and can bring improvements of 2. 5 to 3 BLEU score on the IWSLT 2019 test set, as compared to the result with the given segmentation.
Ranked #2 on Speech-to-Text Translation on MuST-C EN->DE (using extra training data)
no code implementations • 24 Dec 2020 • Marta R. Costa-jussà, Carlos Escolano, Christine Basta, Javier Ferrando, Roser Batlle, Ksenia Kharitonova
Multilingual Neural Machine Translation architectures mainly differ in the amount of sharing modules and parameters among languages.
no code implementations • 2 Nov 2020 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa, Carlos Segura
On the other hand, Multilingual Neural Machine Translation (MultiNMT) approaches rely on higher-quality and more massive data sets.
no code implementations • 29 May 2020 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa, Mikel Artetxe
We propose a modular architecture of language-specific encoder-decoders that constitutes a multilingual machine translation system that can be incrementally extended to new languages without the need for retraining the existing system when adding new languages.
no code implementations • RANLP 2021 • Jordi Armengol-Estapé, Marta R. Costa-jussà, Carlos Escolano
Introducing factors, that is to say, word features such as linguistic information referring to the source tokens, is known to improve the results of neural machine translation systems in certain settings, typically in recurrent architectures.
no code implementations • EACL 2021 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa, Mikel Artetxe
State-of-the-art multilingual machine translation relies on a universal encoder-decoder, which requires retraining the entire system to add new languages.
no code implementations • WS 2019 • Noe Casas, Jos{\'e} A. R. Fonollosa, Carlos Escolano, Christine Basta, Marta R. Costa-juss{\`a}
In this article, we describe the TALP-UPC research group participation in the WMT19 news translation shared task for Kazakh-English.
no code implementations • IJCNLP 2019 • Carlos Escolano, Marta R. Costa-jussà, Elora Lacroux, Pere-Pau Vázquez
In this context, RNN's, CNN's and Transformer have most commonly been used as an encoder-decoder architecture with multiple layers in each module.
no code implementations • ACL 2019 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa
Multilingual Neural Machine Translation approaches are based on the use of task-specific models and the addition of one more language can only be done by retraining the whole system.
no code implementations • 15 May 2019 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa
By adding and forcing this interlingual loss, we are able to train multiple encoders and decoders for each language, sharing a common intermediate representation.
no code implementations • 15 Oct 2018 • Carlos Escolano, Marta R. Costa-jussà, José A. R. Fonollosa
Preliminary results on the WMT 2017 Turkish/English task shows that the proposed architecture is capable of learning a universal language representation and simultaneously training both translation directions with state-of-the-art results.
no code implementations • WS 2018 • Noe Casas, Carlos Escolano, Marta R. Costa-juss{\`a}, Jos{\'e} A. R. Fonollosa
In this article we describe the TALP-UPC research group participation in the WMT18 news shared translation task for Finnish-English and Estonian-English within the multi-lingual subtrack.
no code implementations • WS 2017 • Marta R. Costa-juss{\`a}, Carlos Escolano, Jos{\'e} A. R. Fonollosa
This paper presents experiments comparing character-based and byte-based neural machine translation systems.
no code implementations • 7 Oct 2016 • Marta R. Costa-jussà, Carlos Escolano
In this paper, we propose to de-couple machine translation from morphology generation in order to better deal with the problem.