Bing in production on Turing: deep AI in worldwide-scale search
In one sentence Microsoft announces a Bing-wide production deployment of Turing-NLR (next-gen NLP) models on Azure GPUs, described as the largest search-quality improvement ever.
Searching on a search engine feels simple but is incredibly hard: the system must understand what you actually mean, even when the phrase is poorly written or ambiguous. For years that capability lived on rules and statistics; now Microsoft shifts to deep learning.
They roll out a new generation of their language models across all of Bing, trained on Azure with thousands of GPUs. The practical consequence: Bing starts giving more relevant answers to complex queries, and handles synonyms and intent better.
It's a rare 2020 example of "large AI applied to a service for hundreds of millions of users", and sets the stage for GPT-4's integration into Bing three years later.
Companies
Microsoft
Tools
Turing-NLR, Bing
Tags
Sources