Contributors
Albert Xu, Aniket Jasani, Snigdha Majeti, Arjun Balaganesh, Chandra Bhogadi, Ghazal Kaviani, Greg Krudysz, and Ghassan AlRegib
Vision
AI First application centered on providing first insight into key AI concepts. Each lesson invokes graphical visualization to motivate each concept along with salient corresponding parameters. Users build their conceptual understanding by building a model of the concept, its state parameters, and through corresponding code examples. Users can use their own data and get first-hand experience in applying AI concepts to their problems or interests.
Release 1.0
Linear regression graphical user interface demonstration based on two illustrative datasets. Users review lecture summaries of “linear regression” slides, select 2- or 3-dimensional data, plot linear regression equations, and interact with key parameters. Users have the option to view and edit Python code scripts for dynamic graphical updating.
Description
Assist GT students within classroom tutoring system-driven LLM-enhanced instruction. Initially covering the concept of regression and then moving on to a wider range of machine learning topics. Tutorials with real-time coding environments provide students with a thoroughly immersive and practical learning experience. The instructional materials utilize Python notebooks, then transformed into interactive slide-like modules, and lastly run straight on the high-powered computing nodes. This tool is targeting all campus members regardless of their technical background to receive a guided hands-on learning experience in learning the fundamentals concepts in AI and Machine learning. The tool provides a rich dashboard that captures the user’s progress.
Features
- LLM-empowered tutoring: chat support powered by Mistral or any other open-source models
- Interactive guided lessons: markdown + LaTeX slides along with cells that can be executed
- Managed compute access: running safely on PACE nodes with Python kernel
- Live demonstrations: neural network training and regression visualization are among the demos
- Learning personalization: learner-centric zero-shot summaries
- Ingestion tools: transforming PDFs/PPTs into markdown lessons and RAG sources
Video and Demo
