Authors: Rahul Garg and Mitch Mason, Product Managers, Watson Solutions

Train…Test…Deploy…

Building cognitive Question and Answer applications is now as easy as 1 – 2 – 3 with the IBM Watson Natural Language Classifier and Watson Dialog services. These two APIs have commonly been combined in enterprise applications but we are now making them available to developers as part of the Watson Developer Cloud on the IBM Bluemix platform. With these services you can embed natural language processing, conversation and deep learning to build engaging cognitive applications quickly and easily.

classifier-64Natural Language Classifier (NLC) is designed to enable your application to understand natural language and react accordingly based on the meaning behind the text. It is based on deep learning, a relatively recent set of approaches with similarities to the way that the human brain works. Deep learning algorithms offer state of the art approaches in image and speech recognition, and NLC now applies deep learning technologies to text classification. Read Rob Yates’ recent blog post that provides additional details and context about the service.

We continue to work with key partners and developers to enhance the service, and have seen a variety of different ways in which NLC can be used. Many developers used NLC in the classic sense— to help answer questions from users across different channels such as SMS, Twitter, live chat, email and other mediums:

  • One Partner is converting incoming voice calls to text using IBM Speech to Text and NLC to understand the intent of the call to route it to the appropriate department.
  • We are also observing developers commonly use NLC in conjunction with other Watson services. One of the most popular use cases is combining NLC and Dialog to create innovative and engaging applications.

iDAvatars is using NLC to build a Virtual Health Assistant to communicate with their users. See a video here.

“iDAvatars is thrilled to be an early adopter of Natural Language Classifier, an IBM Watson API which allows our users to retrieve information from Watson based on the intent of their question rather than the precise nature in which they asked the question.  NLC uses an efficient method of tagging information so that, using natural language, the user is able to text or speak into our app and quickly receive the information they need, either in conversation or text format.  NLC has reduced our cost and effort by over 60% compared with prior methods and we continue to expand its use with each new client.” – Norrie Daroga, CEO iDAvatars

Dialog_Icon-60pxThe Watson Dialog service combines the convenience of an automated system with the personal touch and intelligence that is usually found only in interactions between people. The service was specifically designed to enhance applications by providing an automated system that can have a conversation with people through natural language. When speaking to someone, you often expect the person to have a sense of humor, to understand colloquialisms and context.

The Dialog service can now help an application have the same type of conversation with users. One of the most common use cases for the Dialog service is building a virtual agent app to have a conversation that:

  • Shows personality
  • Recognizes informal expressions
  • Uses context clues that are gathered during the conversation to understand conversation
  • Disambiguates between alternative meanings to help users complete a task.

“There is an enormous opportunity with IBM’s Dialog to create engaging characters that consumers can interact with. We have created simulations that bring to life movie and TV characters online, and also medical simulations for the health industry. By using Dialog, we can immerse audiences in ways that have never been achieved before, pioneering the future of storytelling.” – Guy Gadney, Group Executive Director, The Project Factory

The Watson Dialog service uses a set of building blocks to construct the logic for a conversational system that knows when to present information to users and when to request information from users. Developers can populate the service with a rich set of grammars that enable virtual agents to use words and phrases that might be expected only in real conversation. The service even comes with ready-made templates based on conversational analysis, derived from an academically recognized study of human interactions in everyday life. These templates help to create high-quality interactive systems more quickly than starting from scratch.

The Watson Dialog Service has been used with many of the Watson Engagement Advisor Enterprise customers as part of their solutions. Deakin University in Australia uses the service to help students navigate around campus and advise them about their academic careers. Watch this video to learn more about how Deakin University uses the Dialog service to improve student life.

With the addition of Dialog to the Watson Developer Cloud portfolio, Welltok is advancing their commercial application by using the service to bring new levels of personalization and engagement to users.

“Watson Dialog enables more interaction on a personal level, which helps us better understand and connect with consumers to drive positive health decisions and behaviors,” – Jeff Margolis, Chairman and CEO, Welltok

There have been a number of virtual agents available for some time, however Watson’s unique combination of NLC and Dialog makes it possible to build a much more sophisticated virtual agent. Using a deep learning model coupled with a Dialog system you could create a virtual agent that rivals first-level customer support representatives and provide a great experience for users. Imagine a website where you need to login and you forgot your password. Currently that experience requires you to go to multiple web pages to reset your password, and sometimes wait for an e-mail. With NLC and Dialog you could create a simple application where, through natural language, a user can reset their password.

RR1

Applications today require a multi-step process for users to change passwords. Using Watson services you can create an application with a very natural flow as you would have with a real agent.

RR2

We have also posted sample code and applications in the Watson App Gallery to help you jumpstart building your application. We are excited to see what you will build with the Watson Dialog and Natural Language Classifier Services on the Bluemix Platform.

______________________________________________

IBM is placing the power of Watson in the hands of developers and an ecosystem of partners, entrepreneurs, tech enthusiasts and students with a growing platform of Watson services (APIs) to create an entirely new class of apps and businesses that make cognitive computing systems the new computing standard.

9 comments on"Combining Natural Language Classifier and Dialog to create engaging applications"

  1. Is there an URL where we can try this dialog application directly?
    The videos are nice to get an idea of the overall capabilities, but having the opportunity of really trying the app live would give a better first hands experience, and help understanding how to use this for our own apps.

  2. Could you please help me how many languages Dialog service is supported ?

    • Currently the Dialog API only supports English, with new language support planned for 2016.

      • Thanks Keely!

        I need to build a app using Dialog service and Natural Classifier as a POC.
        I have successfully build apps for dialog and Natural classifier in nodejs.
        Could you please help me how i can combine these two apps into one?
        or do you have any example of dialog and classifier in nodejs ?

Join The Discussion

Your email address will not be published. Required fields are marked *