Purpose

The purpose of AIBridge is to bridge the gap between computer science and other disciplines. To many, working with AI might seem like an unreachable objective. However, one week is enough to get started. AIBridge provides students basic programming capability in Python and knowledge of object-oriented programming as well as the concepts behind machine learning and how to implement it using a popular toolbox, Scikit-Learn. Students work to complete a personally-defined project using techniques in AI, with data from their own research or with problems supplied by the course. 

This first iteration of this one-week course was hosted in-person at UC Davis June 27–July 1, 2022 for undergraduates, graduate students, and postdoctoral researchers. 

Instructors

AI Bridge Instructor: Professor Xin Liu

Teaching Assistants: Albara Ah Ramli, Houjun Liu, Samuel Ren

Format

5 days with a morning and afternoon session, each with 1.5 hours of lecture followed by a 1.5 hours of labs to practice.

Objective

The goal was to teach students the basics so that they could work out a simple machine learning (ML) problem themselves from the beginning to the end using the Sklearn toolbox. This camp equipped them with the tools and ability to learn more on their own, take more advanced classes, and collaborate more easily.

Therefore, as the name AIBridge suggests, the curriculum was designed to focus on applying ML to real research problems. Algorithm discussion focused on intuition and how to apply that. This bootcamp focused on learning by doing – students practiced what they learned in the lab sessions with TAs and course instructor answering questions on the spot and showing them how to debug. 

Content

The camp started with 1.5 days on Python programming focusing on what is needed for machine learning, specifically sklearn. This included basic syntax, variable types, control loops, logic and operations, list manipulation, functions, and classes.

The next 3.5 days focused on ML using sklearn. The course covered supervised learning, including linear regression, logistic regression, SVM, naive Bayes, decision tree/random forest, and unsupervised learning, kmeans and PCA. Students learned about data preprocessing, including missing data, feature selection, feature engineering, debugging ML outcome, and how to work on a ML project. 

AIBridge concluded with a final project presentation. The final project had two parts: 

Part 1: perform classification and regression using an existing wine dataset. Students presented their results using algorithms that were not covered in the class. 

Part 2: walk through a research project using ML, from why ML is a good approach, how to collect data, potential challenges and bias, to what input/output for ML and what ML algorithms to use. Students shared about potential applications to their own research, with topics from food safety and supplement, to mosquitoes and bird migration.