Inspiration

Education is increasingly going digital, but many of the tools currently available for teaching and learning using AI act like chatbots—one person asks a question, and then a response comes back.

We wanted to create something different.

We aim to develop an AI tutor more like a real teacher. MentorSLM will not solve a problem for the student but will first try to understand what the student already knows, find where the student is having difficulty, explain the concept at the appropriate level, provide them with hints when they fail, and support them in their retries.

The concept that learning is not only about the right answer, inspired us. It's about understanding WHY an answer is right, what it is wrong, and it's about developing expertise over time.

It motivated us to develop MentorSLM, a Small Language Model supported adaptive education platform.


What it does

MentorSLM is an AI-based education platform to offer a personalized and structured learning experience.

The learning journey starts with:

Sign Up / Sign In → Send an email to ensure a secure registration (Email Verification) → Two-Step Verification → Set a profile → Start up learning → Topic Search → Knowledge Assessment → Student Model → Teacher SLM

The platform will first determine how much a student knows before giving a topic for their class.

For instance, when a student selects Python, MentorSLM could assess their knowledge of:

  • Variables
  • Conditions
  • Loops
  • Functions
  • Lists
  • Object-Oriented Programming

Then a personalized student knowledge profile is created.

MentorSLM doesn't start at the top, rather it begins at the point where the student requires assistance.

The Teacher SLM then works together with four main engines:

Teaching Engine

The Teaching Engine presents the concepts to students based on their present level.

It can provide:

  • Step-by-step explanations
  • Examples
  • Hints
  • Simplified explanations Cater for the student's readiness with more advanced explanations

The objective is NOT to give the learner the answer, but to lead the learner to finding it by him or herself.


Practice Engine

At the end of each lesson the student will receive exercises on the Practice Engine.

As students submit answers, it verifies if the answer is correct.

But if the answer is wrong, MentorSLM doesn't just change that answer with the correct answer.

Rather it is sent to Mistake Analyzer.


Mistake Analyzer

The Mistake Analyzer is used to identify what might have been the misconception that the student may have drawn from their answer.

Built With

Share this project:

Updates

Submission history