Events

Teaching computers how to distinguish between different types of animals and flowers is the bare minimum when it comes to machine learning. As a beginner to the field though, understanding how to teach a computer program how to classify is a nuanced problem with a lot of different moving parts -- it can feel like when you started learning how to read cursive. You know how to read individual cursive letters (fundamental programming syntax), but first putting all the letters together to read a paragraph can be incredibly frustrating. The same is true with artificial intelligence (AI) and machine learning (ML).

The field of artificial intelligence and machine learning is fast-growing, but with that comes a lot of high knowledge barriers for beginners. Our project aims to empower learners, from programming hobbyists to industry professionals who are new to the field, with introductory notebooks to walk them through the process of gathering and manipulating data to training and evaluating an ML model. The project contains two components: a React website which acts as a discovery point and hosts a high-level overview of educational content and programming code blocks, and separate interactive Jupyter notebooks that contain modifiable code and explanatory content for an hands-on exploration of handling different types of data and building customized models. Both will feature unique and interactive data visualizations; these will help to demonstrate the type of data that users will be feeding into the models as well as live feedback on the performance of a trained model.

The current content covers different models such as logistic regression, neural networks, and convolutional neural networks as well different types of classification problems like multi-categorical classification, image classification, and sentiment/text classification. 

Although there are a handful of existing resources introducing AI and ML concepts, we have bundled a full experience into this project. The majority of resources only provide some combination of a visualization, a programming tutorial, or concept explanation but not all three. Overall, our team is hoping to provide strong computer science education for an evolving field and help familiarize the concepts for anyone who may be interested. Our project will provide a scaffolding to help users build confidence in exploring AI and ML using simple and accessible datasets. 

This project was completed under the guidance of Dr. Minsuk Kahng, the team's project partner and advisor. 

For any inquiries, you may contact Laura at jianglau@oregonstate.edu.