Seamless: Multilingual Expressive and Streaming Speech Translation

• Published in: arXiv.org 2023-12
• Main Authors: Communication, Seamless, Barrault, Loïc, Yu-An, Chung, Mariano Coria Meglioli, Dale, David, Dong, Ning, Duppenthaler, Mark, Paul-Ambroise Duquenne, Ellis, Brian, Hady Elsahar, Haaheim, Justin, Hoffman, John, Min-Jae Hwang, Inaguma, Hirofumi, Klaiber, Christopher, Kulikov, Ilia, Li, Pengwei, Licht, Daniel, Maillard, Jean, Mavlyutov, Ruslan, Rakotoarison, Alice, Kaushik Ram Sadagopan, Ramakrishnan, Abinesh, Tran, Tuan, Wenzek, Guillaume, Yang, Yilin, Ye, Ethan, Evtimov, Ivan, Fernandez, Pierre, Gao, Cynthia, Hansanti, Prangthip, Kalbassi, Elahe, Kallet, Amanda, Kozhevnikov, Artyom, Gabriel Mejia Gonzalez, Robin San Roman, Touret, Christophe, Wong, Corinne, Wood, Carleigh, Yu, Bokai, Andrews, Pierre, Balioglu, Can, Peng-Jen, Chen, Costa-jussà, Marta R, Elbayad, Maha, Gong, Hongyu, Guzmán, Francisco, Heffernan, Kevin, Jain, Somya, Kao, Justine, Lee, Ann, Ma, Xutai, Mourachko, Alex, Peloquin, Benjamin, Pino, Juan, Popuri, Sravya, Ropers, Christophe, Safiyyah Saleem, Schwenk, Holger, Sun, Anna, Paden Tomasello, Wang, Changhan, Wang, Jeff, Wang, Skyler, Williamson, Mary
• Format: Article
• Language: English
• Subjects: Communication; Human bias; Linguistics; Machine translation; Multilingualism; Speech; Voice communication
• ISSN: 2331-8422

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