End-to-end Speech Translation with Self-supervised Speech Representations

For years speech translation has been faced as concatenation of speech recognition and machine translation. The powerful architectures of deep learning has made end-to-end speech translation feasible. The student will have to use the encoder-decoder architecture based on Transformer to build multili...

Full description

Bibliographic Details
Author: Gallego Olsina, Gerard Ion|||0000-0001-7466-3606
Format: master thesis
Publication Date:2020
Country:España
Institution:Universitat Politècnica de Catalunya (UPC)
Repository:UPCommons. Portal del coneixement obert de la UPC
Language:English
OAI Identifier:oai:upcommons.upc.edu:2117/332513
Online Access:https://hdl.handle.net/2117/332513
Access Level:Open access
Keyword:translation
speech
text
end-to-end
transformer
self-supervision
pase
apc
wav2vec
Description
Summary:For years speech translation has been faced as concatenation of speech recognition and machine translation. The powerful architectures of deep learning has made end-to-end speech translation feasible. The student will have to use the encoder-decoder architecture based on Transformer to build multilingual speech translation systems.