Meta, the company formerly known as Facebook, will this year have the most powerful supercomputer in the world dedicated to artificial intelligence tasks. The machine, known as the AI Research SuperCluster, or RSC, is already up and running, albeit not at full computing power, and will be used to generate machine learning models capable of working in everything from comment moderation to the design of virtual environments.
"We hope RSC will help us build entirely new AI systems that can, for example, power real-time voice translations to large groups of people, each speaking a different language, so they can seamlessly collaborate on a research project." or play an augmented reality game together," explain Kevin Lee and Shubho Sengupta, engineers in charge of the project.
 |
| A Cyberattack |
RSC will be capable of performing trillions of operations per second once it is completed thanks to the use of more than 16,000 graphic processing units (which are used in artificial intelligence environments due to their high parallel calculation capacity). Today the computer has almost a third of those drives.
This calculation capacity, according to the company, is necessary to continue advancing in domains such as artificial vision or language recognition, in which machine learning techniques have generated very useful tools in recent years but in which there are still obstacles. important to overcome.
 |
| Who Has Seen Your Profile |
"We're looking for an infrastructure that can train models with more than 1 billion parameters on data sets as large as an exabyte, which, to give some context, is the equivalent of 36,000 years of high-quality video, " Lee said. Sengupta.
The previous supercomputer that Facebook used to train its AI models is made up of 22,000 Nvidia V100 processing units. The new machine, which will be 20 times faster in computer vision processes, will instead use Nvidia's DGX A100 architecture. For practical purposes, this increased power will allow machine learning models to be trained much more quickly. A model with tens of billions of parameters, for example, can complete training in three weeks, compared to nine weeks with the previous model.
 |
| WhatsApp Chat History |
Once created, these models can be executed very quickly, and this is the technology behind many of the tools that we now consider common within Meta services, such as automatic photo tagging or video transcription.
To train these machine learning algorithms, Meta uses data from Facebook, Instagram or WhatsApp users, but the company claims that it has integrated several mechanisms into RSC to protect user privacy. The data, for example, remains encrypted until it is processed and RSC lacks a direct connection to the network, the information can only be sent or received from the company's data processing centers.
Do You Know What We Have Posted on
0 Comments