Microsoft wants to take traditional search engines to the next level with the open sourcing of SPTAG or Space Partition Tree and Graph.
According to the company, search engines and even intelligent assistants are lacking the ability to understand queries because they are focused on keyword search algorithms. “Keyword search algorithms just fail when people ask a question or take a picture and ask the search engine, ‘What is this?’” Rangan Majumder, group program manager on Microsoft’s Bing search and AI team, said in a blog post.
SPTAG uses artificial intelligence to make web searching feel more natural to the user. Microsoft explained SPTAG is an algorithm that uses deep learning models to search through vectors and deliver relevant results faster. This differs from keyword search because it enables users to search by concept. Microsoft has been using this technology in its Bing search engine in order to better understand users’ search intents.
“Bing processes billions of documents every day, and the idea now is that we can represent these entries as vectors and search through this giant index of 100 billion-plus vectors to find the most related results in 5 milliseconds,” said Jeffrey Zhu, program manager on Microsoft’s Bing team.
Other use cases for SPTAG include ability to identify languages from audio snippets and the ability to provide a more accurate search results from images.
The open-source project will also provide user examples and a video on how to use SPTAG.