ML APPLICATIONS

Powered by AI: Advancing product understanding and building new shopping experiences

Today we’re unveiling advancements of our AI-powered shopping system that leverages state-of-the-art image recognition models to improve the way people buy, sell, and discover items. This work is foundational to transform the way people shop on Facebook.

May 19, 2020

ML APPLICATIONS

Powered by AI: Advancing product understanding and building new shopping experiences

May 19, 2020

RESEARCH

Hateful Memes Challenge and data set for research on harmful multimodal content

May 12, 2020

ML APPLICATIONS

A highly efficient, real-time text-to-speech system deployed on CPUs

May 11, 2020

Blog Posts

Blog

April 20, 2020

RESEARCH

ML APPLICATIONS

Using AI to help health experts address the COVID-19 pandemic

Facebook AI is partnering with academic researchers and other experts to create forecasting models of the spread of COVID-19 in order to help healthcare providers and emergency responders determine how best to allocate resources.

April 20, 2020

Learn More

Blog

April 14, 2020

ML APPLICATIONS

Nevergrad, an evolutionary optimization platform, adds new key features

We have added a range of noteworthy new features to Nevergrad, Facebook AI’s open source Python3 library for derivative-free and evolutionary optimization.

April 14, 2020

Learn More

Blog

April 08, 2020

RESEARCH

New code completeness checklist and reproducibility updates

Today, @paperswithcode introduced its new ML Code Completeness Checklist, building on the AI reproducibility work from FAIR Managing Director Joelle Pineau

April 08, 2020

Learn More

Blog

March 09, 2020

ML APPLICATIONS

OPEN SOURCE

The evolution of deep learning and PyTorch

Watch co-creator of PyTorch, SoumithChintala share the story of how PyTorch was created, how it’s evolved, and what the future holds for the community-driven framework.

March 09, 2020

Learn More

Blog

March 05, 2020

RESEARCH

ML APPLICATIONS

Using integrated ML to deliver low-latency mobile VR graphics

We are sharing details on a new low-latency, power-efficient framework for running machine learning in the rendering pipeline for standalone VR devices that use mobile chipsets.

March 05, 2020

Learn More

Blog

February 28, 2020

RESEARCH

ML APPLICATIONS

Powered by AI: Turning any 2D photo into 3D using convolutional neural nets

We have created a new AI-powered tool that can turn virtually any standard 2D picture into a 3D image.

February 28, 2020

Learn More

Blog

February 27, 2020

RESEARCH

ML APPLICATIONS

Poincaré maps: Hyperbolic embeddings to understand how cells develop

February 27, 2020

Learn More

Blog

February 25, 2020

RESEARCH

COMPUTER VISION

FastMRI leverages adversarial learning to remove image artifacts

As part of the fastMRI initiative, Facebook AI and NYU Langone Health researchers have used a technique known as an orientation adversary to remove artifacts from AI-accelerated scans. This removes a major obstacle blocking deep learning use in MRIs.

February 25, 2020

Learn More

Blog

February 13, 2020

OPEN SOURCE

RESEARCH

Optimizing infrastructure for neural recommendation at scale

To advance the understanding of how to optimize neural recommendation systems, we are sharing insights from how these systems perform in production at scale, as well as relevant performance metrics for evaluation.

February 13, 2020

Learn More

Blog

February 13, 2020

RESEARCH

ML APPLICATIONS

Read to Fight Monsters: Using RL to teach agents to generalize to new settings

A grounded RL problem, Read to Fight Monsters, and an approach called txt2π has agents read and learn complex tasks, enabling them to generalize to new environments.

February 13, 2020

Learn More

Blog

February 06, 2020

NLP

OPEN SOURCE

CCMatrix: A billion-scale bitext data set for training translation models

With 4.5B parallel sentences in 576 language pairs, CCMatrix is the largest data set of high-quality, web-based bitexts for training translation models. Now Facebook AI is sharing tools for other researchers to use this corpus for their work.