Seamless: Multilingual Expressive and Streaming Speech Translation

View on arXiv ← Back to list

Authors: Seamless Communication, Loïc Barrault, Yu-An Chung, Mariano Coria Meglioli, David Dale, Ning Dong, Mark Duppenthaler, Paul-Ambroise Duquenne, Brian Ellis, Hady Elsahar, Justin Haaheim, John Hoffman, Min-Jae Hwang, Hirofumi Inaguma, Christopher Klaiber, Ilia Kulikov, Pengwei Li, Daniel Licht, Jean Maillard, Ruslan Mavlyutov, Alice Rakotoarison, Kaushik Ram Sadagopan, Abinesh Ramakrishnan, Tuan Tran, Guillaume Wenzek, Yilin Yang, Ethan Ye, Ivan Evtimov, Pierre Fernandez, Cynthia Gao, Prangthip Hansanti, Elahe Kalbassi, Amanda Kallet, Artyom Kozhevnikov, Gabriel Mejia Gonzalez, Robin San Roman, Christophe Touret, Corinne Wong, Carleigh Wood, Bokai Yu, Pierre Andrews, Can Balioglu, Peng-Jen Chen, Marta R. Costa-jussà, Maha Elbayad, Hongyu Gong, Francisco Guzmán, Kevin Heffernan, Somya Jain, Justine Kao, Ann Lee, Xutai Ma, Alex Mourachko, Benjamin Peloquin, Juan Pino, Sravya Popuri, Christophe Ropers, Safiyyah Saleem, Holger Schwenk, Anna Sun, Paden Tomasello, Changhan Wang, Jeff Wang, Skyler Wang, Mary Williamson

Published: 2023-12-08 17:18:42+00:00

AI Summary

This paper introduces Seamless, a family of models enabling end-to-end expressive and multilingual speech translation in a streaming fashion. This includes improvements to the SeamlessM4T model (SeamlessM4T v2), a new expressive model (SeamlessExpressive), and a streaming model (SeamlessStreaming), along with responsible AI initiatives for safety and mitigation of biases.

Abstract

Large-scale automatic speech translation systems today lack key features that help machine-mediated communication feel seamless when compared to human-to-human dialogue. In this work, we introduce a family of models that enable end-to-end expressive and multilingual translations in a streaming fashion. First, we contribute an improved version of the massively multilingual and multimodal SeamlessM4T model-SeamlessM4T v2. This newer model, incorporating an updated UnitY2 framework, was trained on more low-resource language data. SeamlessM4T v2 provides the foundation on which our next two models are initiated. SeamlessExpressive enables translation that preserves vocal styles and prosody. Compared to previous efforts in expressive speech research, our work addresses certain underexplored aspects of prosody, such as speech rate and pauses, while also preserving the style of one's voice. As for SeamlessStreaming, our model leverages the Efficient Monotonic Multihead Attention mechanism to generate low-latency target translations without waiting for complete source utterances. As the first of its kind, SeamlessStreaming enables simultaneous speech-to-speech/text translation for multiple source and target languages. To ensure that our models can be used safely and responsibly, we implemented the first known red-teaming effort for multimodal machine translation, a system for the detection and mitigation of added toxicity, a systematic evaluation of gender bias, and an inaudible localized watermarking mechanism designed to dampen the impact of deepfakes. Consequently, we bring major components from SeamlessExpressive and SeamlessStreaming together to form Seamless, the first publicly available system that unlocks expressive cross-lingual communication in real-time. The contributions to this work are publicly released and accessible at https://github.com/facebookresearch/seamless_communication

Key findings

SeamlessM4T v2 achieves state-of-the-art results in various speech and text translation tasks. SeamlessExpressive shows improvements in preserving prosody and vocal style, although some trade-offs with speech clarity are observed. SeamlessStreaming demonstrates simultaneous translation with controlled latency, showing varying performance across language resources and families.

Approach

The approach uses a cascaded architecture. SeamlessM4T v2, improved with the UnitY2 framework, provides a foundation for multilingual translation. SeamlessExpressive preserves vocal styles and prosody, addressing aspects like speech rate and pauses. SeamlessStreaming uses Efficient Monotonic Multihead Attention for low-latency translation.

Datasets

SeamlessAlign (automatically aligned data for 76 languages), mExpresso (multilingual expressive speech data), mDRAL (multilingual dialogues re-enacted across languages), Fleurs, CoVoST2, CVSS, Multilingual HolisticBias, MuTox (speech toxicity data), various publicly available datasets.

Model(s)

SeamlessM4T v2 (with UnitY2), SeamlessExpressive (with Prosody UnitY2 and PRETSSEL), SeamlessStreaming (with EMMA), w2v-BERT 2.0, Conformer, HiFi-GAN, PRETSSEL, Unit Voicebox, AutoPCP, SeamlessWM watermarking system, MuTox toxicity detector, Whisper ASR, NLLB, XLS-R.

Author countries

France, Taiwan, Argentina, United Kingdom

← Previous