Skip to content
conftrace_
2020 ACL ACL 2020

DiDi Labs’ End-to-end System for the IWSLT 2020 Offline Speech TranslationTask

Abstract

AbstractThis paper describes the system that was submitted by DiDi Labs to the offline speech translation task for IWSLT 2020. We trained an end-to-end system that translates audio from English TED talks to German text, without producing intermediate English text. We use the S-Transformer architecture and train using the MuSTC dataset. We also describe several additional experiments that were attempted, but did not yield improved results.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Natural Language Processing and Speech & Audio

Authors

Arkady Arkhangorodsky, Yiqi Huang, Amittai Axelrod

Topics

Artificial Intelligence > Core AI > Multimodal Learning Deep Learning > Architectures > Transformers Natural Language Processing > Applications > Machine Translation Speech & Audio > Recognition > Speech Translation

Keywords

neural machine translation end-to-end learning speech translation speech-to-text translation end-to-end speech translation end-to-end system audio translation ted talk offline translation must-c dataset
Download PDF

More from ACL 2020

All ACL 2020 papers →
Start-Before-End and End-to-End: Neural Speech Translation by AppTek and RWTH Aachen University
KIT’s IWSLT 2020 SLT Translation System
End-to-End Simultaneous Translation System for IWSLT2020 Using Modality Agnostic Meta-Learning
End-to-End Offline Speech Translation System for IWSLT 2020 using Modality Agnostic Meta-Learning
End-to-End Speech-Translation with Knowledge Distillation: FBK@IWSLT2020
SRPOL’s System for the IWSLT 2020 End-to-End Speech Translation Task