Engineering Blog

Blog posts tagged 'Artificial Intelligence'

Under the hood: Photo Search

Posted last Monday

Facebook’s Photo Search team applied machine learning techniques to better understand what’s in an image as well as improve the search and retrieval process. Read more...

ParlAI: A new software platform for dialog research

Posted about 2 weeks ago

ParlAI is a one-stop shop for dialog research, where researchers can submit new tasks and training algorithms to a single, shared repository. Read more...

A novel approach to neural machine translation

Posted about 3 weeks ago

FAIR's convolutional neural network approach for language translation achieves state-of-the-art accuracy at 9x the speed of recurrent neural systems. Read more...

Expanded fastText library now fits on smaller-memory devices

Posted about a month ago
blog post · Artificial Intelligence · Research · C++

fastText now has 294 languages and fits on smartphones and small computers like Raspberry Pi. Read more...

Lauren RuganiTechnology Communications at Facebook

Facebook's new tech at F8 2017

Posted about a month ago

Day two of F8 2017 was packed with announcements about Facebook's big technology bets across connectivity, video and virtual reality, and AI. Read more...

Enhancing high-resolution 360 streaming with view prediction

Posted about a month ago

We gave our system the ability to intuit the most interesting parts of a video, to help prioritize where to concentrate the pixels in a stream. Read more...

Faiss: A library for efficient similarity search

Posted about 2 months ago

Vector representation allows for fast, large-scale image searches where traditional key/value queries fall short. Read more...

Evaluating boosted decision trees for billions of users

Posted about 2 months ago

Efficiency improvements resulted in evaluating larger models in the same amount of time and with the same CPU resources. Read more...

Kevin LeeEngineering

Introducing Big Basin: Our next-generation AI hardware

Posted about 3 months ago

With Big Basin, we can train machine learning models that are 30 percent larger, which enables our researchers and engineers to move more quickly in developing increasingly complex AI models. Read more...

Jeff ReynarEngineering at Facebook

Machine Learning @Scale 2017 recap

Posted about 4 months ago

Facebook New York hosted more than 300 people to discuss large-scale applied machine learning systems. Read more...

Register now for F8 2017!

Posted about 4 months ago

Apply now for a spot at Facebook's annual developer conference. Read more...

Building scalable systems to understand content

Posted about 4 months ago

Pushing computer vision to understand images at the pixel level helps our systems describe photos for the visually impaired and provide better search results for visual content. Read more...

Facebook Open Source 2016 year in review

Posted about 5 months ago

Over the past few years, Facebook's Open Source program has grown into one of the largest and most active portfolios in the industry. Read more...

Miki FriedmannEngineering

Made in NY: The engineering behind social recommendations

Posted about 6 months ago

The product incorporates machine learning and client-side caching to identify relevant posts and dynamically update attachments as new recommendations are added. Read more...

Artificial intelligence, revealed

Posted about 6 months ago

The heads of Facebook AI Research and Applied Machine Learning teams break down some of the commonly used concepts in artificial intelligence. Read more...

Tobias TieckeEngineering at Facebook

Open population datasets and open challenges

Posted about 6 months ago

Sharing data about how people are aggregated around the world can help solve challenges such as connectivity, infrastructure planning, humanitarian aid, and disaster response. Read more...

Mike SchroepferChief Technology Officer at Facebook

Accelerating innovation and powering new experiences with AI

Posted about 7 months ago

Research and engineering in connectivity, virtual reality, and artificial intelligence will help us make progress toward opening the world to everyone. Read more...

Delivering real-time AI in the palm of your hand

Posted about 7 months ago

Caffe2go is a new deep learning platform that can — for the first time — capture, analyze, and process pixels in real time, directly on a mobile device. Read more...

Building an efficient neural language model over a billion words

Posted about 7 months ago

Facebook AI Research (FAIR) releases a library for recurrent neural networks as well as a novel softmax function approximation tailored for GPUs. Read more...

Lauren RuganiTechnology Communications at Facebook

Highlights from @Scale 2016

Posted about 9 months ago
blog post · Data · Mobile · @Scale · Tooling · Video · Open Source · Artificial Intelligence

Engineers representing hundreds of companies gathered to discuss the challenges and opportunities of building apps and systems at scale. Read more...

Piotr DollarResearch Scientist at Facebook

Segmenting and refining images with SharpMask

Posted about 9 months ago

New detection technologies will move us toward a more precise understanding of images. Read more...

FAIR open-sources fastText

Posted about 9 months ago

The new open source fastText library quickly builds state-of-the-art text classifiers that scale to billions of words. Read more...

Erin GreenEngineering

@Scale 2016 lineup announced!

Posted about 10 months ago

Registration for the 2016 @Scale conference is now open. Read more...

Building a better way to write posts in multiple languages

Posted about 11 months ago

The multilingual composer lets people compose a single post in multiple languages, enabling diverse audiences to more easily connect and share. Read more...

Lighting the way to deep machine learning

Posted about 11 months ago

Open source Torchnet helps researchers and developers build rapid and reusable prototypes of learning systems in Torch. Read more...

A path to unsupervised learning through adversarial networks

Posted about 11 months ago

Adversarial networks provide a strong algorithmic framework for building unsupervised learning models that incorporate properties such as common sense, getting us one step further in the quest to build smarter AI. Read more...

Introducing DeepText: Facebook's text understanding engine

Posted about 12 months ago
blog post · Infra · Data · Artificial Intelligence · Research · News Feed

DeepText can understand with near-human accuracy the textual content of several thousand posts per second, spanning more than 20 languages. Read more...

Keep Updated

Stay up-to-date via RSS with the latest open source project releases from Facebook, news from our Engineering teams, and upcoming events.

Subscribe
Facebook © 2017