Microsoft made its Space Partition Tree And Graph (SPTAG) algorithm available to all as an open source GitHub project. This algorithm is used in their own Bing search engine. The algorithm uses vector search and deep learning models, which provides results based on search intent.
The intent of the algorithm is designed to allow users to search through billions of pieces of information (known as vectors) in milliseconds. That means that, in theory, relevant results are delivered more quickly.
This isn’t a new concept, but it’s made possible through vectorizing data, which is the process of assigning a numerical representation to a word, image pixel or other data point. If the meaning of a piece of data is captured in this way and applied with deep learning models to associate it with other terms, Microsoft can begin understanding and representing search intent. This means that results that match what the users wants.
By making the algorithm available to all, Microsoft is shifting from a closed ecosystem to one that’s accessible and inviting.It was uploaded to Github, which shows its efforts to court the developer community.
With the algorithm being open to all, this means that developers can use the vector search technology to build their own search engines or help improve it by submitting updates. The Bing team is predicting that it will be used for enterprise or consumer-facing applications.