Today, we would like to introduce AI Research SuperCluster (RSC). It is one of the fastest AI supercomputers running today and we believe it will be the fastest in the world if fully built in mid-2022. AI can now perform tasks such as translating text between languages and identifying potentially malicious content, but to develop the next generation of AI will require powerful supercomputers capable of quintillions of operations per second.
RSC will help Meta AI researchers build better AI models that can learn from trillions of cases; Work in hundreds of different languages; easily parse text, images and videos together; develop new augmented reality tools and more. Ultimately, the work done with RSC will lay the groundwork for the technology for the metaverse, a next-generation computing platform in which AI-powered applications and products will play an important role.
Why We Need AI at This Scale
Since 2013, we have made significant strides in the field of AI, including self-supervised learning, where algorithms learn from numerous untagged examples and transformers, allowing AI models to infer more effectively by focusing on specific areas of their input. . Fully realizing the benefits of advanced AI will require training on much larger and more complex models in areas as diverse as vision, speech, and language, especially for critical use cases such as malicious content detection. In early 2020, we decided that the best way to accelerate progress would be to develop a new computing infrastructure, RSC.
Using RSC to Build for the Metaverse
With RSC, you can train models that use multimodal signals faster and determine if an action, sound, or image is harmful or benign. This research will not only help keep users of the current service safe, but will also help in the future to build for the Metaverse. As the RSC moves to the next phase, we plan to make it bigger and more powerful as we begin to lay the foundation for the Metaverse.
