MindCare

• 

  Screenshot of MindCare demo

Inspiration

I noticed the stigma surrounding mental health and how mental health is a problem firsthand during my time as a freshman in college and I wanted to create a product that could help others. I was also inspired by my dad as he had previously worked at a company called CareMessage, where they use an SMS-based system to be able to fulfill health needs of underpriviledged people who may not have access to the latest smart phone. This inspired me to create MindCare as I wanted to tackle the mental health crisis. I decided against using an SMS-based system and used a web-based system as well as Watson Assistant to create a chatbot because I wanted MindCare to be convenient and personalized to each user.

What it does

MindCare is an AI chatbot that users can talk to get immediate help with mental health issues.

How we built it

The MindCare chatbot is hosted on IBM Cloud and uses Watson Assistant AI, which is powered by machine learning and language models to understand the near infinite ways that a user can say the same thing. For example, a user can say they feel good, great, happy, fantastic, etc. While these are technically different responses, Watson Assistant will be able to detect they all have the same meaning. MindCare will be available initially only on the web.

Challenges we ran into

Getting the dialogue to work as intended was difficult, as I often encountered problems with the chatbot saying a different dialogue than intended. Specifically, I wanted the chatbot to say “I hope I could help! If you ever need to talk again, just let me know how you're feeling.” after recommending the user resources, but it either always said it even if the user provides invalid input or it never said it at all. Eventually, I managed to figure out how to use conditionals to check if one of the valid inputs had been used.

Accomplishments that we're proud of

I’m proud of getting the opportunity to learn to create an AI chatbot with IBM Watson Assistant that can help many people.

What we learned

I learned how IBM Watson Assistant uses machine learning, language models, and deep learning to analyze the user’s message and accurately judge the intent of their message to be able to deliver the correct response.

What's next for MindCare

I plan to add more dialogue options and resources as well as making the responses more human and helpful. I’m also planning to add some features to the chatbot such as sentiment analysis and a physician alerting system in case the user is feeling depressed or suicidal. I also plan on adding mobile app support once MindCare grows and develops more.

Built With

• html
• ibm-cloud
• ibm-watson
• watson-assistant-ai
• watson-assistant-api