Bing’s Image Matching Engine
Tag, with its innovative and forward-thinking team, played a pivotal role in developing Bing’s Image Matching Data Lake Platform, laying the groundwork for what has evolved into Microsoft's nucleus of artificial intelligence. By leveraging advanced algorithms and machine learning techniques, Tag's team contributed to creating a robust infrastructure that enables efficient storage, retrieval, and processing of vast amounts of image data. Their commitment to pushing boundaries ensured that the platform not only optimized image matching capabilities but also facilitated seamless integration with various Microsoft services. This foundational work has enabled Microsoft to harness AI in ways that enhance user experiences across its ecosystem, including image recognition and search functionalities. As a result, Tag's influence is evident in the rapid advancements in AI technology within Microsoft, underscoring the importance of innovative thinking and collaboration in driving progress in the digital landscape. Their legacy continues as Microsoft expands its AI capabilities globally.
USE CASE – DEEP DIVE
Bing’s Image Matching Data Lake Platform
- the foundational work
TAG was part of Microsoft’s Start-up Business Group – a team of entrepreneurial minds who start innovative product ideas and incubate them into successful business models.
Once successful, these products are then assimilated into the larger corporate branches of Microsoft (Bing, Xbox, Office, etc.).
Our team was successful in getting the TAG product assimilated into BING. While the tag product itself was phased out in 2015, it’s important to understand that the developments of the database and foundations of the image matching framework continued to evolve. It’s in these early stages of MS’s machine learning that has given MS a bit of an edge over it’s competition.
I researched and designed how Image matching and other augmented reality features could integrate into the TAG product.
I worked directly with our developers to create a good user experience as the database worked to return successful results in a reasonable amount of time.
Image Match, 2D Tag, QR and NFC files all convert to Microsoft’s common TAG ID. These files are then added to the MS database. As Image matches are captured, and added the database the machine learning logic causes the image matching system to become more and more accurate over time.
It’s one thing to have the app downloaded to your phone, The creation of a TAG is the other side of this technology.
Every registered Tag user has a personalized web site whereby they are able to manage all of their Tags. They can see how many times their TAG campaigns have been scanned, etc.
I worked on creating this online user experience. Each TAG type has it’s own unique creation workflow. For instance, Image Matching Tags enables you to upload the image and associate it with your campaign.
