Microsoft's Bringing Cognitive Services to a Container Near You
If there is one thing that Microsoft is good at doing, it’s shoving its services into every conceivable use-case possible. I don’t mean this in a negative light, the company is good at adapting its platforms to where customers need solutions; Microsoft is announcing this week that it is adding container support to some of its cognitive services.
Starting today, Microsoft will begin offering, in the preview, key phrase extraction, language detection, sentiment analysis, face/emotion detection and OCR recognition container support. The company says that they will be adding more features in the future.
The goal of supporting containers with these services is to get the compute closer to the edge, where real-time analysis is needed. While you can off-load these features to other compute-stacks like the cloud, by having this type of functionality available where the data is collected, it can reduce the latency from detection to action which is a big win in time-sensitive environments.
Further, containers make it easier to architect one solution and deploy it to multiple iterations of the edge and at the same time, deploy new AI models at their own pace without significant overhead. The goal is to make it easier to work with AI which is one of Microsoft’s big bets for the next decade and by containerizing solutions like cognitive services, this aligns to the company’s objective to bring AI to where the data is collected.
In addition to these services being offered in containers, Microsoft is also enabling logo detection with its custom vision service. This tool allows customers to build their own logo detectors to help search for and locate logos in their media libraries.
One of the challenges with AI is not how the services can be used but when to trust them. We are still very much in the infancy of the development and deployment of AI in the corporate environments. While Microsoft and every other company operating in this space continues to make a big claim about how AI will change the world, the adoption and integration into working environments continues to be slow.