Too socially anxious to discuss your personal feelings and experiences with a friend? Play MindTac.
Helping a loved one who has dementia recollect their fondest memories? Play MindTac.
Finding it hard to start a personal conversation on your first date? Play MindTac.
How it works
Two players engage in a fun game of Tic-Tac-Toe, and each move challenges them to answer a thought-provoking personal question. By transforming personal reflection into a fun game, MindTac provides a safe space for players to connect and grow emotionally.
๐ก Inspiration - Many college students who suffer from poor mental health often struggle with open communication and emotional expression. In an effort to reduce social anxiety, MindTac helps by providing a fun, gamified environment where they can explore their feelings, connect with peers, and foster positive thinking through thought-provoking conversations. MindTac isnโt limited to college students - it can also benefit older adults struggling with depression or anxiety, as well as individuals facing cognitive impairments.
โญ๏ธ Our Curiosity: Can we train AI to simulate emotional intelligence? MindTac is powered by AI and Natural Language Processing. Specifically, the app uses a Generative Transformer Model to generate empathetic text prompts, and performs Sentiment Analysis to deliver adaptive feedback based on userโs responses.
How We Built It
- Retrained a generative transformer model โ๏ธ (GPT-2) on our custom dataset using Hugging Face ๐ค to improve its fine-tuning for our natural language processing task.
- Implemented functions to create contextually relevant questions and correct grammar errors in the text generated by the transformer model.
- Developed a ๐ถ๏ธ Flask API endpoint to facilitate prompt generation requests in our application with minimal delay.
- Integrated a ๐ ๐ sentiment analysis library from JavaScript to evaluate user input and calculate sentiment polarity scores for adaptive feedback.
- Designed and implemented an aesthetic and functional user interface featuring ๐ฎ Tic Tac Toe game logic using React, HTML, and CSS.
Challenges we ran into
- Tried to use TextBlob from Python but could not fetch, process and return user input
- Faced errors while working on state management and data flow of React components
- Cross-Platform collaboration issues between our team due to OS specific errors
Accomplishments that we're proud of
We successfully implemented the features we brainstormed during the design phase of our app, and it was incredibly rewarding to see our ideas gradually come to life in a tangible application.
What we learned
Throughout our development process, we learned 2 important things:
- If something doesn't work, look at it from a new perspective.
- Persevere through every unending error.
What's next for MindTac ๐
We want to make the feedback generated by our app more context-aware of the user's sentiments, in order to make personalized suggestions to help improve mood and well-being. We also hope to explore ways to build better, powerful AI to support, not replace human emotional intelligence.



Log in or sign up for Devpost to join the conversation.