AI has been a focus for Microsoft for quite sometime, so it should be expected that they’d announce a supercomputer built for the future of AI work.
The supercomputer is of the top five publicly disclosed in the world, making new infrastructure available in Azure to train extremely large artificial intelligence models. The platform will enable a next generation of AI models whose potential benefits extend far beyond narrow advances in one type of AI model.
Microsoft Chief Technical Officer Kevin Scott says, “This is about being able to do a hundred exciting things in natural language processing at once and a hundred exciting things in computer vision, and when you start to see combinations of these perceptual domains, you’re going to have new applications that are hard to even imagine right now,”
Historically machine learning has been one to one, to put it even more simply: Shows computer image of a cat, machine identifies the cat. The new class of models has proven that some of the basic tasks can be performed better by a single massive model. This would tackle basic tasks like translating between languages, recognizing objects, read text to identify key points or speech recognition. This model is able to better absorb the nuance of language, grammar, context and concepts. This would result in its ability to summarize a lengthy speech, moderate live chats, find relevant passages in legal documents or even generate code from scouring GitHub.
Earlier this year Microsoft released to researchers, as part of a company wide AI at Scale initiative, the largest publicly available AI language model in the world, the Microsoft Turing model for natural language generation. This is the world’s largest publicly available language AI model with 17 billion parameters. At BUILD 2020, Microsoft announced they will soon begin open sourcing its Microsoft Turing models, as well as recipes for training them in Azure Machine Learning.
The goal, Microsoft says, is to make its large AI models, training optimization tools and supercomputing resources available through Azure AI services and GitHub so developers, data scientists and business customers can easily leverage the power of AI at Scale.
The supercomputer developed for OpenAI is a single system with more than 285,000 CPU cores, 10,000 GPUs and 400 gigabits per second of network connectivity for each GPU server. Compared with other machines listed on the TOP500 supercomputers in the world, it ranks in the top five.
Running Azure, the supercomputer is for customers who want to pursue their AI ambitions but don’t require a dedicated supercomputer.
How the Turning Model and the new supercomputer are helping Microsoft’s products
The Microsoft Turing model for natural language understanding, has rolled out across all of Microsoft’s products.
In Bing, has significantly advanced caption generation and question answering, improving answers in some markets by up to 125%. Microsoft is also exploring ways to search Bing by understanding what’s inside images and video.
In Office, the smart find feature enabling easier searches in Word and the Key Insights feature that extracts important sentences.
In Outlook, Suggested replies which automatically generates possible responses to an email.
Dynamics 365 Sales Insights also uses it to suggest actions to a seller based on interactions with customers.
Microsoft’s BUILD 2020 was held digitally this year, as a result there are a ton of resources available online.