Edu.Aid: Personalized AI Educational platform for children !

We propose an education system powered by AI and Large language models (LLMS) that tailors learning to each individual child's needs and mental well-being. It aims to unlock the potential of kids.

Comment 1

Inspiration

My inspiration is that we : Dr Omneya Attallah, Professor with the department of Electronics and communications Engineering at the at Arab academy for science technology and maritime transport, and myself, Eng.Sara Zayan, teaching assistant with the same department encountered a lot of students with different learning abilities, most of students with learning abilities are very creative and very smart but they struggle with the regular learning system. In Egypt, we have limited resources that enable personalizing the learning system. Furthermore, staff members are not equipped with the tools and experience to deal with these students. That's why we decided to create a simple system to identify their mental health ability and design a personalized learning system and pace for each student.

What it does

This prototype uses kaggle datasets from adhd and depression to classify between adhd and depression using only LLM and advanced prompting .LLM used is Gemini.

How we built it

This prototype was built using Gemini LLM , advanced prompting and langchain

Challenges we ran into

challenges are the gathering of dataset , textual data on adhd, depression , dyslexia is really limited

Accomplishments that we're proud of

we are proud to achieve 93.3% accuracy on only 15 samples using only prompting techniques !!

What we learned

we learnt that careful prompt techniques can do very complex tasks . Careful prompting principles may produce wonderful results

What's next for Edu.Aid: Personalized AI Educational platform for children!

The next step is to :

  1. Make the code more modular .
  2. Add more classes like dyslexia , anxiety etc.
  3. Perform cross-sectional study and gather data from multiple institutions and wider range of age and different background
  4. Validate the model in real world scenarios in different schools , universities or colleges and determine if it improved the learning experience of these students
  5. Use Gemini1.5 when I have access to improve results especially when the context length of the prompt increases .

Built With

Share this project:

Updates