MSBuild tour 2017 quick recap on AI

 One of the primary focuses at recent #Microsoft event was on making applications and the cloud intelligence. In these sessions, it was about some of the latest advancements in intelligent services that you can make use. For brief overview of Cognitive Services here

It is open to consume for every developer. These services use images and videos as input which understands by linking to real-time and translates into your own languages. Speech Recognitions, programming, knowledge understanding, Q/A maker features are built in features. 

Computer vision, semantic and language understanding intelligent services are interesting to analyze and provide information.

More Info on LUIS: https://azure.microsoft.com/en-us/services/cognitive-services/language-understanding-intelligent-service/

Easy to be customized into own language and domain. Custom vision service and search service will use your own data and calculate recognition algorithms to make simple with better results.

For example: let's say you use below sample image

Sample rest API: https://westcentralus.api.cognitive.microsoft.com/vision/v1.0/analyze?visualFeatures=Categories,Description,Color&language=en

Sample response on the right side: 

For custom vision service, you can create project with set of image and respective tags associated with images, this gives you most accurate information like below

More info on Computer vision: https://azure.microsoft.com/en-us/services/cognitive-services/computer-vision/

  

Custom Vision service:

https://azure.microsoft.com/en-us/services/cognitive-services/custom-vision-service/

How to use the Prediction API in custom vision service

If you have an image URL:

https://southcentralus.api.cognitive.microsoft.com/customvision/v1.0/Prediction/{ID}/url?iterationId={iteration_Id}

Set Prediction-Key Header to : {Key_ID}

Set Content-Type Header to : application/json

Set Body to : {"Url": "<image url>"}

If you have an image file:

https://southcentralus.api.cognitive.microsoft.com/customvision/v1.0/Prediction/{ID}/image?iterationId={iteration_Id}

Set Prediction-Key Header to : {Key_ID}

Set Content-Type Header to : application/octet-stream

Set Body to : <image file>

Sample results:

Custom speech service:

You can have customised speaking styles like kids voice, echo, recordings which help to recognised speech as outputs.

https://azure.microsoft.com/en-us/services/cognitive-services/custom-speech-service/

Next is "Bot": Yes, What is it???.

In a simple terminology, it’s just restful service which takes request as query and sends the response back using intelligence, which can be connected to different conversion channels/canvas.

  

Bots and its common patterns, we see when people are building them, new channels they can be connected to, and great ways to surface them.

Flow of typical a bot service and conversation with connectors

If you are building a bot, it is safe to assume that you are expecting users to use it. It is also safe to assume that you are hoping that users will prefer the bot experience over alternative experiences like apps, websites, phone calls, and other means of addressing their needs. 

Complete Handoff and supervised handoff for solving complex problems and more cases to improve.

Caller to BOT to Supervisor (Custom solution)

Bot to Web:

Few other patterns like “Apps” bots, proactive messages are available.

There are few factors which you need to consider when you are using bot

·        Does the bot easily solve the user’s problem with the minimum number of steps?

·        Does the bot solve the user’s problem better/easier/faster than any of the alternative experiences?

·        Does the bot run on the devices and platforms the user cares about?

·        Is the bot discoverable? Do the users naturally know what to do when using it?

Regardless of the type of application, you're creating (bot, website, or app), make the user experience a top priority.

Bot Architecture:

Do you want to create a bot? refer here

Cards: Twitter cards and Facebook cards which you are normally aware when you share your web page to different social networks.

how you can save yourself from implementing a different UI for every potential channel is with Adaptive Cards??

Bot with different containers:

Do you want to build you own AI?? See here

I will be adding more information on other sessions in coming articles.