CSE 190D / 291J: Special Topics: Fairness, bias, and transparency in Machine Learning (Winter 2025)


Instructor	Julian McAuley
Room	Peterson Hall, 108
Days & times	9:30-10:50am, Tuesdays & Thursdays
Office hours	posted on Piazza

[piazza] [gradescope] [twitch] [podcast]

Description

This course is devoted to fairness, bias, and transparency in machine learning. After taking this course, you will be able to understand the main sources of bias and unfairness in machine learning systems, and deploy strategies to mitigate these biases. You will also understand the related notions of accountability and transparency in Machine Learning, allowing for the development of systems that are more trustworthy.

Schedule

module	week (approx)	slides	(w/annotations)	workbook
		intro and outline
1	1-2	regression and classification	(w/ annotations)	workbook 1 and datasets
2	3-4	intro to bias and fairness	(w/ annotations)	workbook 2
3	4-5	fairness and bias interventions	(w/ annotations)	workbook 3
4	6-7	interpretable and explainable AI	(w/ annotations)	workbook 4
5	8-9	fairness and bias in application domains	(w/ annotations)
	10	project presentations? (if time allows)

Assignments

All work must be completed on your own unless otherwise specified.
Please submit your work to Gradescope. See Piazza for entry code.
Late homework submissions will accepted up to five days late, subject to a reduction of 2 points per day.

Homework	50%	Assignments	50%
├ Homework 1	10%	├ Assignment 1	25%
├ Homework 2	10%	├ Assignment 2	25%
├ Homework 3	10%
├ Homework 4	10%
├ Homework 5	10%