MindMate

Inspiration

Mental health is becoming an increasingly important topic, especially amongst students. Sometimes when you’re facing a mental health problem it’s hard to find resources or even just open up to someone. MindMate was created so that students would have a platform that could assist them in getting the help they need, whenever they need.

What it does

This mental health chatbot is designed to detect the emotions you’re feeling based on the input you give it. Then, it will give you practical and useful resources such as therapists near you, exercises that to make you feel better, media recommendations, and any other resources that the chatbot thinks would help you.

How we built it

Take input from a user
Use fine tuned AI model to identify the emotions they are feeling based on input
Based on the emotions detected as well as what resources the user wants, use Google Custom Search API to identify practical resources
Put the identified resources into a message that the chatbot displays back to the user

Challenges we ran into

Time was a big constraint as some parts of our project were rather ambitious. However, we were able to make sure we met the deadline with a good quality deliverable.

Another problem we had was finding good data to fine tune our data model. This was something that we had to ensure we took the time to gather properly. We were able to get good data in the end, and it helped us make our chatbot unique and helpful.

Accomplishments that we're proud of

Finetune a pre trained model
Figured out how to use the output from the AI in order to help people
Integrated the output of the model with practical resources that people could use

What we learned

how to fine tune an AI model

What's next for MindMate

add customized entertainment suggestions such as movies and music based on their interests as a resource option
use location information to find local resources such as programs at a library near the user