Microsoft is open sourcing the AI-powered software algorithms it uses to power Bing search services, on GitHub. Called Space Partition Tree And Graph (SPTAG), the algorithm makes Bing an advanced search engine using deep learning models. These models enable Bing to deliver millions of pieces of information, called vectors, in milliseconds.
By open sourcing SPTAG, Microsoft wants to empower developers to use these algorithms for their own use cases where end-users perform searches among large data troves.
In the past, web search was very easy. The end users just expected some relevant results from the search engines. But things have changed today. The same users now can perform searches by dropping an image in the search box.
They can ask digital assistants like Cortana to search for things. Plus, when they ask a question, they expect the search engine to return the exact answer, rather than showing a list of pages with answers. This is making things difficult for traditional search engines.
Advancements in AI is powering the search engines like Bing to address such challenges. Microsoft uses SPTAG algorithm for Bing to power it with vector search.
“Vector search makes it easier to search by concept rather than keyword. For example, if a user types in “How tall is the tower in Paris?” Bing can return a natural language result telling the user the Eiffel Tower is 1,063 feet, even though the word “Eiffel” never appeared in the search query and the word “tall” never appears in the result,” explained Microsoft in a blog post.
The Bing team mentioned that open sourcing the SPTAG algorithm will help enterprises developers to power their applications with capabilities that can identify a language being spoken based on an audio snippet. It can also make apps capable of identifying pictures.
Also read: Microsoft rolls out new AI capabilities in Azure for developers and enterprises
“Even a couple seconds for a search can make an app unusable,” noted Rangan Majumder, group program manager on Microsoft’s Bing search and AI team.
“We’ve only started to explore what’s really possible around vector search at this depth.”
