ELITR Non-Native Speech Translation at IWSLT 2020

TL;DR

This paper presents ELITR's speech translation systems for IWSLT 2020, including offline and real-time ASR and SLT, emphasizing non-native speech handling and system selection strategies.

Contribution

It introduces new end-to-end and hybrid ASR systems trained on non-native speech and discusses system selection for non-native speech translation at IWSLT 2020.

Findings

Developed a new end-to-end general ASR system.

Trained a hybrid ASR on non-native speech data.

Compared multiple system candidates for non-native speech translation.

Abstract

This paper is an ELITR system submission for the non-native speech translation task at IWSLT 2020. We describe systems for offline ASR, real-time ASR, and our cascaded approach to offline SLT and real-time SLT. We select our primary candidates from a pool of pre-existing systems, develop a new end-to-end general ASR system, and a hybrid ASR trained on non-native speech. The provided small validation set prevents us from carrying out a complex validation, but we submit all the unselected candidates for contrastive evaluation on the test set.