The latest from Meta AI

Natural Language Processing

Bringing the world closer together with a foundational multimodal model for speech translation

SeamlessM4T provides high-quality translation, allowing people from different linguistic communities to communicate effortlessly through speech and text.

August 22, 2023

Meta Connect

Expanding reality, today and tomorrow

Join us on September 27 and 28 for a virtual event focused on AI and virtual, mixed and augmented realities.

Learn more

Get the latest from Meta AI in your inbox

Blog Posts

Blog

May 18, 2023

Research

ML Applications

Pursuing groundbreaking scale and accelerating research using Meta’s Research SuperCluster

Meta’s RSC is one of the fastest AI supercomputers in the world. Today, we’re sharing how we are leveraging RSC’s power to accelerate AI research at scale and deliver impactful results in a fraction of the time.

May 18, 2023

Learn More

Blog

May 18, 2023

Hardware

ML Applications

MTIA v1: Meta’s first-generation AI inference accelerator

In 2020, we initiated the Meta Training and Inference Accelerator (MTIA) family of chips to support our evolving AI workloads, starting with an inference accelerator ASIC for deep learning recommendation models (DLRMs).

May 18, 2023

Learn More

Blog

May 18, 2023

Hardware

MSVP: Meta’s first ASIC for video transcoding

The Meta Scalable Video Processor (MSVP) will support video on demand and live streaming, as well as generative AI and AR/VR content.

May 18, 2023

Learn More

Blog

May 18, 2023

Hardware

ML Applications

Reimagining Meta’s infrastructure for the AI age

Meta is executing on an ambitious plan to build the next generation of its infrastructure backbone – specifically for AI. This includes our first custom chip for running AI models, a new AI-optimized data center design, and phase 2 of our 16,000 GPU supercomputer for AI research.

May 18, 2023

Learn More

Blog

May 11, 2023

ML Applications

New AI advancements drive Meta’s ads system performance and efficiency

Using a high-capacity architecture allows our ads system to more broadly and deeply understand new concepts and relationships in data and benefits advertisers through joint optimization of a large number of goals.

May 11, 2023

Learn More

Blog

May 09, 2023

Research

Computer Vision

ImageBind: Holistic AI learning across six modalities

ImageBind is the first AI model capable of binding information from six modalities.

May 09, 2023

Learn More

Blog

April 25, 2023

Research

The Self-Supervised Learning Cookbook

Meta AI has released a new “Cookbook of Self-Supervised Learning," which provides AI researchers and practitioners with a practical guide on how to navigate SSL methods. This is part of our efforts to lower the barrier and help democratize access to SSL research.

April 25, 2023

Learn More

Blog

April 17, 2023

Computer Vision

DINOv2: State-of-the-art computer vision models with self-supervised learning

Today, we are open-sourcing DINOv2, the first method for training computer vision models that uses self-supervised learning to achieve results that match or surpass the standard approach used in the field.

April 17, 2023

Learn More

Blog

April 13, 2023

Developer Tools

Research

A new, unique AI dataset for animating amateur drawings

With this dataset and animation code, we believe that the domain of amateur drawings can inspire a new generation of creators with its expressive and accessible possibilities.

April 13, 2023

Learn More

Blog

April 05, 2023

Research

Computer Vision

Introducing Segment Anything: Working toward the first foundation model for image segmentation

We're releasing the Segment Anything Model (SAM) — a step toward the first foundation model for image segmentation — and the SA-1B dataset.

April 05, 2023

Learn More

Blog

March 31, 2023

Research

ML Applications

Robots that learn from human videos and simulated interactions

Training robots in the wild is slow, expensive, and dangerous. FAIR has developed an artificial visual cortex (VC-1) from egocentric videos and visuomotor skills from simulated environments that can form the foundation for embodied intelligence. VC-1 achieves impressive results on 17 different sensorimotor tasks in virtual environments, and the visuomotor skills deployed on the Spot robot achieve near-perfect performance on the task of navigating and fetching objects in indoor environments.

March 31, 2023

Learn More

Blog

March 24, 2023

ML Applications

How Meta measures the management of its AI ecosystem

AI development ecosystems are increasingly complex and challenging to maintain. We're sharing details on our work to develop highly efficient systems to build, serve and improve AI models for production uses.

March 24, 2023

Learn More

Prev Next