insanelab.com insanelab.com

August 8, 2018 - Cognitive Computing

How You Can Use Microsoft’s Cognitive Computing in Your Business

by Antoni Żółciak
More by this author

Artificial Intelligence (AI) is the hot new thing. If you want to remain on the cutting edge and outperform your competitors, you need to develop intelligent applications that interact with humans in the same way as a live person would. You are expected to work with apps that understand native language, identify and analyze trends, learn from experience, and provide data for evidence-based decisions.

Artificial Intelligence is the New Electricity
Andrew Ng

And since Insane Lab is working mainly with Microsoft technologies, we decided to cover them in-depth. The Redmond giant has developed a set of machine learning algorithms to address major AI challenges; these algorithms are called Cognitive Services.

Cognitive Services break down areas of AI into specific components so that developers can use them in their own applications. Hence, Cognitive Services are a set of application programming interfaces (APIs). An API is basically an interface for software that allows it to interact with other software.

Before we proceed any further, however, let us remind you of what cognitive computing is in general.

Cognitive Computing. What Is It, Really?

Peter Sommer, IBM

To us, the end users, technology is becoming transparent. There is nothing more “normal” than our smartphones and computers. Cognitive computing makes the machine more intelligent, more friendly. More… Human. It surely is not nor it will be an emotional or spiritual experience, but it is more than interacting with devices or objects that are not smart in any way.

Apple’s Siri is a good example. If you ask Siri for an hour, she will not respond with just the result of your question. She will ask you about your mood, as well. “Oh, it’s 1 PM already. How about that lunch, Michael?”.

Cognitive computing is a term that describes technology platforms that are based on some form of AI and signal processing. These platforms include:

  • Machine learning
  • Artificial intelligence
  • Natural language processing
  • Speech recognition
  • Object recognition

And more. There is no strict definition for cognitive computing in either academia or industry, yet.

Cognitive computing is like giving eyes and a brain to your computer

Some people would say that cognitive computing represents the third era of computing: from tabulating sums to programmable systems to cognitive systems. Key enterprise players implement cognitive technologies in a variety of different ways.

IBM Watson, for example, is a “question answering” machine that combines artificial intelligence and analytical software.

Google’s DeepMind has developed an AI that has managed to learn how to walk, run, jump, and climb without any prior guidance. And it looks just as strange as it sounds.

Long story short, cognitive computing is an automation-based next step in computing. It will try to set a benchmark, to try and reach the level of the human brain. The only question remains: how close can we get?

Build Intelligent Applications with Microsoft Cognitive Services

Microsoft Cognitive Services APIs allow developers to access the ultimate vision, speech, language, and knowledge technologies to build intelligent applications to enhance businesses. They simplify a wide array of intelligent tasks. Microsoft’s idea was to enable companies to benefit from years of research and utilize the most up-to-date intelligence technologies with only a few lines of code added to one’s app.

Microsoft’s Cognitive Service APIs integrate into any language and will work on any platform – whether it is iOS, Android or Windows. Moreover, each one of these experiences is being constantly improved through machine learning.

Microsoft’s Cognitive Service APIs are grouped into five categories.

Cognitive Service Vision APIs

The Vision APIs include five groups of tasks that focus on analyzing content, images, and videos to smartly identify, caption, and moderate pictures.

  • Computer Vision classifies images, recognizes handwriting and recognizes optional characters, giving you actionable data for your business. Additionally, Computer Vision recognizes scenes, activities, celebrities, and landmarks within images.
  • Video Indexer analyzes videos within your app, including object, activity, scene, and face recognition.
  • Custom Vision offers customizable image recognition.
  • Face provides face, person, and emotion detection in images as well as the detection of similar faces.
  • Content Moderator automatically screens content for profane, offensive, or explicit content and offers tools for feedback from human moderators.

Cognitive Service Speech APIs

The Speech APIs include four groups of tasks that are tools used to integrate speech processing into any app. When used together these tools create natural sounding speech from text and easily recognizes most speech.

  • Speech to Text offers automatic speech recognition and transcription. This API is also customizable and also provides solutions for other vocabularies and accents.
  • Text to Speech is an automatic and customizable text to speech tool.
  • Speaker Recognition offers speaker identification and verification for your software application.
  • Speech Translation is real-time, automated, and customizable speech translation.

Cognitive Service Language APIs

The Language APIs include four groups of tasks that interpret intent and process unstructured text. The Language API group also includes the Content Moderator API to moderate content for explicit content.

  • Text Analytics extracts key phrases and analyzes the sentiment behind those phrases. This tool also provides named entity recognition.
  • Bing Spell Check provides specific language-based spell check as well as contextual spell checking.
  • Language Understanding is a tool that understands language within context.
  • Translator Text helps you to deal with other languages by providing automatic language detection and translation. You can also customize translation for your own needs.

Cognitive Service Knowledge API

The Knowledge API has one task group, QnA Maker, which provides three services that utilize or create rich resources that can be integrated into apps and services:

  • Extracts questions and answers from unstructured text.
  • Creates a knowledge base from collections of questions and answers.
  • Matches knowledge bases to keywords.

Cognitive Service Search APIs

Broadly speaking, the search APIs apply machine learning to Bing web searches with eight different tools that enable apps to have a web-scale ad-free search engine, so you can find exactly what you want on the internet.

  • Bing Web Search is an advertisement-free and safe web search that use location services.
  • Bing Visual Search identifies and classifies image while acquiring knowledge from images and identifying similar images.
  • Bing Custom Search is used to create a custom search engine with ad-free search results.
  • Bing Entity Search recognizes and classifies named entities, while storing knowledge from these searches.
  • Bing Video Search identifies topics and trends with an ad-free video search.
  • Bing News Search provides trending topics and offers ad-free news search results.
  • Bing Image Search is an ad-free image search.
  • Bing Autosuggest auto-completes any searches.

The tools and APIs are there. Your business can be enhanced in a matter of weeks, and you do not have to develop complicated, cutting-edge software from scratch. Essentially, you just use the results of years of research and development. Working with a competent software team can help to plan a strategy and implement these solutions in accordance with your business model.

This collection of cloud-hosted APIs that let add AI capabilities for vision, speech, language, knowledge and search into applications, across devices and platforms such as iOS, Android and Windows have been used by more than a million developers.

But it is easy to throw big promises around and not deliver. Please find some specific use cases of Microsoft Cognitive Services below:

  • Provide extra internal or external security for your business by requiring visual identity verification. You can do this through the Face API by requiring a selfie for smart authentication.
  • Create a bot for your website to help answer consumer or client questions. The bot can answer broad and general questions with the Speech, Language, and Knowledge APIs and escalate to a live person if necessary.
  • Use the custom Speech service for customer service lines, building video games, or anything else you can imagine. This creates a pleasant voice interaction that reduces errors from your clients or consumers.
  • Sort through images to find happy, sad, or angry expressions, giving you the perfect picture for your web content.
  • Get detailed insights about your customers with Computer Vision, so you can understand your clients or customers better, attract new business, and retain existing business.
  • Hone in on a strong brand identity by delivering targeted messages and engaging content based on the content analysis and topic and trend identification.

And this is just the beginning.

Are you interested in implementing elements of cognitive computing in your business?

What is your challenge?

Tell us with any means provided. We'd love to hear from you!