Natural Language Processing & Speech

Breaking down language barriers

We want to break down language barriers across the world for everyone to understand and communicate with anyone—no matter what language you speak.

Our natural language processing and speech researchers focus on the interaction between people and computers using human languages, both in diverse written and spoken forms, to remove the barrier of language from the ability to communicate. The challenges are immense, from the billions of users on Facebook across the 6,000 languages in the world—some even without a writing system, to the informal tone, slang and typos people use on Facebook. Our researchers work on the complex problems that span deep learning/neural networks, natural language processing, language identification, text normalization, word sense disambiguation, and machine learning, to break down the problems, and build and deploy robust language translation solutions.

“Connecting the world, regardless of where you live or what language you speak, is our grand challenge. Our teams of researchers are giving people the ability to do just that with innovate new solutions that break down language barriers and make connecting with anyone, anywhere possible.” Benoit Dumoulin, Engineering Manager

Latest Publications

All Publications

Iterative Answer Prediction with Pointer-Augmented Multimodal Transformers for TextVQA

Ronghang Hu, Amanpreet Singh, Trevor Darrell, Marcus Rohrbach

CVPR - June 14, 2020

An Empirical Study of Transformer-Based Neural Language Model Adaptation

Ke Li, Zhe Liu, Tianxiao Shen, Hongzhao Huang, Fuchun Peng, Daniel Povey, Sanjeev Khudanpur

ICASSP - May 9, 2020

Spatial Attention for Far-Field Speech Recognition with Deep Beamforming Neural Networks

Weipeng He, Lu Lu, Biqiao Zhang, Jay Mahadeokar, Kaustubh Kalgaonkar, Christian Fuegen

ICASSP - May 8, 2020

SkinAugment: Auto-Encoding Speaker Conversions for Automatic Speech Translation

Arya D. McCarthy, Liezl Puzon, Juan Pino

ICASSP - May 7, 2020

Multilingual Composer Demo

0:38 | January 1, 2016

Downloads & Projects

View all Downloads & Projects

The bAbI Project is organized towards the goal of automatic text understanding and reasoning.

FastText is a library for text representation and classification.

Stack RNN is a project gathering the code from the paper Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets by Armand Joulin and Tomas Mikolov.

Careers at Facebook Research
Want to solve some of the most challenging technology problems?