Elizabeth Salesky

Featured Co-authors

Research

Pixel Representations for Multilingual Translation and Data-efficient Cross-lingual Transfer 05/23/2023

We introduce and demonstrate how to effectively train multilingual machi...

A Holistic Cascade System, benchmark, and Human Evaluation Protocol for Expressive Speech-to-Speech Translation 01/25/2023

Expressive speech-to-speech translation (S2ST) aims to transfer prosodic...

Language Modelling with Pixels 07/14/2022

Language models are defined over a finite set of inputs, which creates a...

BibleTTS: a large, high-fidelity, multilingual, and uniquely African speech corpus 07/07/2022

BibleTTS is a large, high-quality, open speech dataset for ten languages...

UniMorph 4.0: Universal Morphology 05/07/2022

The Universal Morphology (UniMorph) project is a collaborative effort pr...

Between words and characters: A Brief History of Open-Vocabulary Modeling and Tokenization in NLP 12/20/2021

What are the units of text that we want to model? From bytes to multi-wo...

Assessing Evaluation Metrics for Speech-to-Speech Translation 10/26/2021

Speech-to-speech translation combines machine translation with speech sy...

A surprisal–duration trade-off across and within the world's languages 09/30/2021

While there exist scores of natural languages, each with its unique feat...

SIGTYP 2021 Shared Task: Robust Spoken Language Identification 06/07/2021

While language identification is a fundamental speech and language proce...

Robust Open-Vocabulary Translation from Visual Text Representations 04/16/2021

Machine translation models have discrete vocabularies and commonly use s...

The Multilingual TEDx Corpus for Speech Recognition and Translation 02/02/2021

We present the Multilingual TEDx corpus, built to support speech recogni...

SIGTYP 2020 Shared Task: Prediction of Typological Features 10/16/2020

Typological knowledge bases (KBs) such as WALS (Dryer and Haspelmath, 20...

SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection 06/20/2020

A broad goal in natural language processing (NLP) is to develop a system...

A Corpus for Large-Scale Phonetic Typology 05/28/2020

A major hurdle in data-driven research on typology is having sufficient ...

Phone Features Improve Speech Translation 05/27/2020

End-to-end models for speech translation (ST) more tightly couple speech...

Relative Positional Encoding for Speech Recognition and Direct Translation 05/20/2020

Transformer models are powerful sequence-to-sequence architectures that ...

Generalized Entropy Regularization or: There's Nothing Special about Label Smoothing 05/02/2020

Prior work has explored directly regularizing the output distributions o...

CMU-01 at the SIGMORPHON 2019 Shared Task on Crosslinguality and Context in Morphology 07/23/2019

This paper presents the submission by the CMU-01 team to the SIGMORPHON ...

Exploring Phoneme-Level Speech Representations for End-to-End Speech Translation 06/04/2019

Previous work on end-to-end translation from speech has primarily used f...

Fluent Translations from Disfluent Speech in End-to-End Speech Translation 06/03/2019

Spoken language translation applications for speech suffer due to conver...

Towards Fluent Translations from Disfluent Speech 11/07/2018

When translating from speech, special consideration for conversational s...

Optimizing Segmentation Granularity for Neural Machine Translation 10/19/2018

In neural machine translation (NMT), it is has become standard to transl...