Machine Learning for Science & Society

Machine Learning for Science & Society

  • Spring 2022

  • time: MW 3:00-4:15pm

  • Professor: Sarah Brown

  • course number: CSC 592: Topics in Computer Science

  • Credits: 4

  • Location: Tyler Hall 108

In this class, we will address the challenges in applying machine learning to scientific research and in high stakes social contexts. On the science side, we will examine the role of ML in research, in particular how it works within knowledge production and how to evaluate ML in line with domain norms. On the social side, we will consider how to ensure ML-based algorithmic decision making systems uphold social values, with a focus on fairness. While these two applications are distinct, many of the challenges translate into common technical problems. Some of the common challenges include:

  • missing data

  • noisy or missing labels

  • multiple objectives

We will look at a range of strategies for identifying and mitigating these problems including:

  • robust evaluation

  • model inspection

  • explanations

  • interpretable models

Format

This will be a synchronous course offered in person.

The course will involve:

  • reading and evaluating ML research papers

  • facilitating and participating in class discussions of the papers

  • producing a replication, demo, or illustration of one concept covered for a broader audience

  • completing a project using ML in a scientific or social domain

  • writing a CS conference style (short & concise) final paper on their project

graduate students are encouraged to do a project related to their research

Prerequisites

To be successful in this class students should have:

  • past experience with machine

  • basic programming skills

  • familiarity with concepts in probability, linear algebra, and calculus that appear in ML

varying skill in these topics is ok, but a general understanding of the basic ideas is important.

Complete this Google form to request a permission number from Professor Brown to enroll in this course. Note that you must be enrolled at URI to take this course and be logged into your URI google account to view that form.