Introduction to Azure Cognitive Services
By Agustín Goñi, Business Analyst
What is Azure Cognitive Services?
In a nutshell, it is a group of tools that are designed to enhance applications across platforms and endow them with some extended capabilities, aimed at mirroring or understanding human interaction, divining meaning and intention from data, providing users with a more helpful and nuanced experience, and making sophisticated interactions available to end users.
The aim behind Cognitive Services, as declared by Harry Shum, Executive Vice President at Microsoft Research, is to imbue end-user apps with many of the lessons learned in the field of Artificial Intelligence, and to brand them with a human side, or at least, a more human-like behavior.
How does it work?
Cognitive Services is comprised of several different technologies spanning five separate scopes: vision, speech, language, knowledge and search. They are all exposed as REST APIs, which means that you can take advantage of them by writing clients for any platform. All you need is a security key (and you can start by obtaining a free one to create demos or POCs here). All your interactions with these tools consist of invoking the resource you want through a specific URI, and passing your security key along with your desired headers in your request.
We can divide the different APIs that are part of Cognitive Services in two broad categories. On one hand, you have the more established and well known technologies such as Bing Search. On the other hand, you have the cutting-edge, off-the-wall stuff like LUIS (Language Understanding Intelligent Service), Emotion API, Knowledge Exploration and Translator (a real time speech and text translator, that is).
Cognitive Services is an evolving platform. Some of these APIs are still in beta, and most or all of them are growing and changing. The current full set is as follows:
We won’t be going into any details concerning each API here. We’ll do that in future articles. For now, let us note that these tools are interesting and powerful in and of themselves, but it is their use in tandem that can truly leverage the resources they contain. A good example of this are the different kinds of searches available in the Search category. They spring from the resources already offered by Bing (though they are not exactly the same), so you might think there is nothing very “novel” about them. But if you combine them with, say, face recognition, or language understanding, you can create very powerful functionalities.
Microsoft’s AI Strategy
Over the last year or so Microsoft has been publishing several small apps that take advantage of various AI technologies, including those contained in Cognitive Services. Many of these apps may seem trivial, unimportant, even anticlimactic when you evaluate them as standalone efforts. Here are some examples, in case you’re curious:
But if you look at the big picture you can get a sense of a larger structure, a proposed strategy centered upon the capabilities that Artificial Intelligence can bring to the table for developers. This has been the main focus of Microsoft’s drive for technology innovation, as expressed by their CEO, Satya Nadella, and echoed in technical presentations everywhere.
Cognitive Services in the future
It is impossible at this point to predict the future reach of this technology. As developers and users get more comfortable with the capabilities it offers, they will come up with new and innovating ways to piece them together. Cognitive Services will lead to a cognitive expansion, as it were. It will expand the minds of user, who will demand more sophistication in their apps, and the minds of developers, who will learn to harness and unleash the power buzzing at their fingertips. It is impossible to foresee where this process will lead, but one thing is for sure. We will only witness the true potential of this technology after this cognitive evolution has taken place. For the moment, we have to wait and see.