MemoMe

• 

  Logo

Inspiration

Over 55 million people worldwide suffer from some form of memory impairment. Memory impairment hampers people’s mental health and independence. Hiring a full-time caregiver costs upwards of 2300 USD per month which is not feasible for most. We wanted to make users more independent and ease their lives by mitigating the hardships that come with memory impairment.

What it does

Personalized information and daily tasks are input into MemoMe by a loved one or a helper. MemoMe then reminds the user to do their daily tasks, tracks their progress for them and allows them to ask dynamic questions throughout. A helper can monitor the daily progress of their patient and is alerted if they need assistance.

How we built it

We used React for the front end and Flask for the backend. We used MongoDB Compass to store user data. We used various APIs such as Speech to Text and Text to Speech services for easy interaction. We used a Hugging face GPT-J LLM Space for chatbot interaction. ## Challenges we ran into There were several challenges we successfully overcame: 1. Developing the backend to categorize input data and parse output data 2. Managing chatbot hallucinations due to incorrectly fed data 3. Dealing with the inherent limitations of an open source LLM ## Accomplishments that we're proud of We were able to establish a plan quickly and execute We were also flexible to technical barriers presented We delegated work to play on each other's strengths

What we learned

We learned about using LLMs in specific applications, using MongoDB for real-world projects, and speech-to-text and text-to-speech model APIs.

What's next for MemoMe

Upgrading the LLM to harness a better memory and larger data inputs, improving entire UX, and finding compatible stand-alone devices like monitors, watches, earbuds, etc.

Built With

• css3
• flask
• mongodb
• python
• react.js