As artificial intelligence grows, the realm of what is possible for communication between a human and a machine grows as well. Machines are inching closer and closer to conversations that could seem indistinguishable from human dialogue. One of the biggest advances in this field is Watson, IBM's computer system designed to answer questions posed to it in natural language.
To put it simply, Watson "listens" to what you say or type, understands what you're trying to do, and quickly offers the most relevant solution for your needs. Watson lets people take the effort out of small questions so they can focus on the important, big ones.
We want to see how you use Watson to create a chatbot that can dialogue with a human subject.
Part 1: Building a Chatbot Application
Start by clearly defining the audience, purpose and desired outcomes of the chatbot. Be sure to clearly state how your chatbot addresses a user challenge or problem. Include other new or different situations or business processes where your bot could be extended. Identify settings where your bot would it be helpful. For example, how could it be used to make someone's day or job better?
- Register for a free IBM Bluemix account and and apply a student promotion code. Qualifying students can get a promotion code which extends access to additional Watson services for 6 months at onthehub.com/ibm. Follow the instructions provided at checkout to apply your unique promotion code.
- Learn about the IBM Watson Conversation service. Explore various demos and workspaces in the documentation and build your conversation logic.
- Create a chatbot app using the Watson Conversation service which can handle at least 7 rounds of dialogue without losing context.
- Upload your Watson Conversation workspace to the Bot Asset Exchange with #mindsumochallenge in the description to showcase your work in the community and help others build their bots.
Part 2: Upgrading to a Virtual Agent
Can you improve your Chatbot application to accept speech input from users in addition to text? And to recognize user tone and emotions?
- Check out IBM Watson's Speech to Text and Text to Speech services. (If you have not already applied a student promotion code from step 1 above you must do so now to access to these services at no charge.)
- Consider incorporating the Tone Analyzer service as well. If you want to connect the tone analyzer to the conversation service, you can download a sample package and modify the app.js file accordingly. The process is documented in the README of this package.
- Design the user experience however you think is best - just make sure it is intuitive and logically addresses the user experience/needs.
- Deploy the chatbot app as publicly accessible on IBM Bluemix platform. You must use the Continuous Delivery: Toolchains functions and configure GitHub as your source code tool integration.
-
Submissions will be graded on the following criteria:
- Meets Deliverables (based on Watson Conversation, at least 7 rounds of dialogue, additional Watson services, deployed on Bluemix, source code posted on github)
- Creativity / Originality
- Application User Experience
- Proposal Clarity and Completeness
will share $1,000
will share $400
will share $200
| $500.00 |
|
Daniel Frick Hochschule Muenchen | |
| $500.00 |
|
Joshua Jolly University of California, Merced | |
| $400.00 |
|
Ranjan Biswas, MCP, CSPO George Washington University | |
| $66.70 |
|
John Hutcherson North Carolina State University | |
| $66.70 |
|
Aparna Siva SantaClara University | |
| $66.70 |
|
Sidharth V AU |