Posts

Humans might lump chess and poker into a single “games” category, but for an AI there’s a huge difference between so-called perfect-information games like chess and imperfect-information games like Texas Hold’em. Today we introduce ReBeL, an algorithm that extends the RL+Search approach (which has helped AI master perfect-information games) to poker, Liar’s Dice, and more.

https://ai.facebook.com/…/rebel-a-general-game-playing-ai-…/

Creating broadly accessible development practices and resources helps to accelerate innovation in AI. We’re proud to be a founding member of MLCommons, an open engineering consortium that is uniting industry leaders to develop and share best practices, benchmarks and metrics across the industry. https://mlcommons.org/…

Image may contain: text that says 'ML Commons'
See All
Videos
We’re releasing fairmotion, a library to help AI researchers use motion capture data for computer graphics and robotics. It provides tools to load, process and visualize motion, and demonstrates its utility with example tasks. We’ve used the library in our work presented at SIGGRAPH 2020 on controlling diverse behaviors for physically simulated characters. Get the code and paper: http://github.com/facebookresearch/fairmotion https://research.fb.com/publications/a-scalable-approach-to-control-diverse-behaviors-for-physically-simulated-characters/
Image
351
16
AI_Integrity_Rio_animation_V2.mp4
Image
163
9
Introducing Opacus: A high-speed library for training PyTorch models with differential privacy
Image
172
5
See All
Photos
Image may contain: text that says 'ML Commons'
No photo description available.
Image may contain: text that says 'New model GAN GANoG of GANs Incremental learning Extract frames Production classifier Manual Manualreview review Extract Extractfaces faces Create new syntheticfaces faces'
See All