Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components

Raul Vazquez; Michele Boggia; Alessandro Raganato; Niki A. Loppi; Stig-Arne Grönroos; Jörg Tiedemann

Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components

Raul Vazquez , Michele Boggia, Alessandro Raganato, Niki A. Loppi, Stig-Arne Grönroos, Jörg Tiedemann

Forskningsoutput: Kapitel i bok/rapport/konferenshandling › Konferensbidrag › Vetenskaplig › Peer review

Sammanfattning

We give an update on the Found in Translation (FoTran) project, focusing on the study of emerging language-agnostic representations from neural machine translation (NMT). We describe our attention-bridge model, a modular NMT model which connects language-specific components through a shared network layer. Our latest implementation supports distributed training over many nodes and GPUs in order to substantially scale up the number of languages that can be included in a modern neural translation architecture.

Originalspråk	engelska
Titel på värdpublikation	Proceedings of the 23rd Annual Conference of the European Association for Machine Translation
Redaktörer	Helana Moniz, Lieve Macken, Andrew Rufener, et al.
Antal sidor	2
Utgivningsort	Geneva
Förlag	European Association for Machine Translation
Utgivningsdatum	2022
Sidor	311-312
ISBN (elektroniskt)	9789464597622
Status	Publicerad - 2022
MoE-publikationstyp	A4 Artikel i en konferenspublikation
Evenemang	Annual Conference of the European Association for Machine Translation - Ghent, Belgien Varaktighet: 1 juni 2022 → 3 juni 2022 Konferensnummer: 23 https://eamt2022.com

Vetenskapsgrenar

6121 Språkvetenskaper
113 Data- och informationsvetenskap

Åtkomst av dokument

2022.eamt-1.45Slutlig publicerad version, 223 KBLicens: CC BY-NC-SA

https://aclanthology.org/2022.eamt-1.45Licens: CC BY-NC-SA

FoTran: Found in Translation - Natural Language Understanding with Cross-Lingual Grounding
Attieh, J., Aulamo, M., Boggia, M., Celikkanat, H., De Gibert Bonet, O., Grönroos, S., Mickus, T., Raganato, A., Scherrer, Y., Sjöblom, E. I., Talman, A., Vazquez , R., Virpioja, S. P., Yli-Jyrä, A., Zosa, E., Tiedemann, J., Celikkanat, H., Raganato, A., Silfverberg, M., Sulubacak, U. & Vazquez , R.
01/09/2018 → 29/02/2024
Projekt: EU Horizon 2020: European Research Council: Consolidator Grant (H2020-ERC-COG)

Citera det här

Vazquez , R., Boggia, M., Raganato, A., Loppi, N. A., Grönroos, S-A., & Tiedemann, J. (2022). Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components. I H. Moniz, L. Macken, & A. Rufener, et al. (Red.), Proceedings of the 23rd Annual Conference of the European Association for Machine Translation (s. 311-312). European Association for Machine Translation. https://aclanthology.org/2022.eamt-1.45

Vazquez , Raul ; Boggia, Michele ; Raganato, Alessandro et al. / Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components. Proceedings of the 23rd Annual Conference of the European Association for Machine Translation. redaktör / Helana Moniz ; Lieve Macken ; Andrew Rufener, et al. Geneva : European Association for Machine Translation, 2022. s. 311-312

@inproceedings{3edfb254460443608c323d5a03040262,

title = "Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components",

abstract = "We give an update on the Found in Translation (FoTran) project, focusing on the study of emerging language-agnostic representations from neural machine translation (NMT). We describe our attention-bridge model, a modular NMT model which connects language-specific components through a shared network layer. Our latest implementation supports distributed training over many nodes and GPUs in order to substantially scale up the number of languages that can be included in a modern neural translation architecture.",

keywords = "6121 Languages, 113 Computer and information sciences",

author = "Raul Vazquez and Michele Boggia and Alessandro Raganato and Loppi, {Niki A.} and Stig-Arne Gr{\"o}nroos and J{\"o}rg Tiedemann",

year = "2022",

language = "English",

pages = "311--312",

editor = "Helana Moniz and Lieve Macken and {Rufener, et al.}, Andrew",

booktitle = "Proceedings of the 23rd Annual Conference of the European Association for Machine Translation",

publisher = "European Association for Machine Translation",

address = "Switzerland",

note = "Annual Conference of the European Association for Machine Translation ; Conference date: 01-06-2022 Through 03-06-2022",

url = "https://eamt2022.com",

}

Vazquez , R , Boggia, M , Raganato, A, Loppi, NA, Grönroos, S-A & Tiedemann, J 2022, Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components. i H Moniz, L Macken & A Rufener, et al. (red), Proceedings of the 23rd Annual Conference of the European Association for Machine Translation. European Association for Machine Translation, Geneva, s. 311-312, Annual Conference of the European Association for Machine Translation, Ghent, Belgien, 01/06/2022. <https://aclanthology.org/2022.eamt-1.45>

Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components. / Vazquez , Raul ; Boggia, Michele ; Raganato, Alessandro et al.

Proceedings of the 23rd Annual Conference of the European Association for Machine Translation. red. / Helana Moniz; Lieve Macken; Andrew Rufener, et al. Geneva : European Association for Machine Translation, 2022. s. 311-312.

Forskningsoutput: Kapitel i bok/rapport/konferenshandling › Konferensbidrag › Vetenskaplig › Peer review

TY - GEN

T1 - Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components

AU - Vazquez , Raul

AU - Boggia, Michele

AU - Raganato, Alessandro

AU - Loppi, Niki A.

AU - Grönroos, Stig-Arne

AU - Tiedemann, Jörg

N1 - Conference code: 23

PY - 2022

Y1 - 2022

N2 - We give an update on the Found in Translation (FoTran) project, focusing on the study of emerging language-agnostic representations from neural machine translation (NMT). We describe our attention-bridge model, a modular NMT model which connects language-specific components through a shared network layer. Our latest implementation supports distributed training over many nodes and GPUs in order to substantially scale up the number of languages that can be included in a modern neural translation architecture.

AB - We give an update on the Found in Translation (FoTran) project, focusing on the study of emerging language-agnostic representations from neural machine translation (NMT). We describe our attention-bridge model, a modular NMT model which connects language-specific components through a shared network layer. Our latest implementation supports distributed training over many nodes and GPUs in order to substantially scale up the number of languages that can be included in a modern neural translation architecture.

KW - 6121 Languages

KW - 113 Computer and information sciences

M3 - Conference contribution

SP - 311

EP - 312

BT - Proceedings of the 23rd Annual Conference of the European Association for Machine Translation

A2 - Moniz, Helana

A2 - Macken, Lieve

A2 - Rufener, et al., Andrew

PB - European Association for Machine Translation

CY - Geneva

T2 - Annual Conference of the European Association for Machine Translation

Y2 - 1 June 2022 through 3 June 2022

ER -

Vazquez R , Boggia M , Raganato A, Loppi NA, Grönroos S-A , Tiedemann J. Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components. I Moniz H, Macken L, Rufener, et al. A, redaktörer, Proceedings of the 23rd Annual Conference of the European Association for Machine Translation. Geneva: European Association for Machine Translation. 2022. s. 311-312

Latest Development in the FoTran Project – Scaling Up Language Coverage in Neural Machine Translation Using Distributed Training with Language-Specific Components

Sammanfattning

Vetenskapsgrenar

Åtkomst av dokument

Projekt

FoTran: Found in Translation - Natural Language Understanding with Cross-Lingual Grounding

Citera det här