Machine learning models are used far and wide, from criminal justice to health to content moderation. In practice, machine learning can often replicate and amplify biases against marginalized groups. How do we identify and measure these biases? What are different metrics used in fairness audits of machine learning models? This 2-hr workshop reviews critical and technical components for FairML. By the end of the workshop, participants will be armed with tools and theory to articulate fairness concerns and evaluate machine learning models.

Computational Text Analysis Working Group (CTAWG)

Title: Privacy and data ethics in text data mining contexts

Abstract: As data becomes available on a massive scale, questions of privacy and data ethics become more prevalent in various research contexts. In this talk, we will address and discuss some common concerns and issues that come up in this space and how to practically deal with them.

Overview

Participants in this workshop will learn about some of the issues surrounding the collection of health statistics, and will also learn about authoritative sources of health statistics and data. We will look at tools that let you create custom tables of vital statistics (birth, death, etc.), disease statistics, health behavior statistics, and more. The focus will be on U.S. statistics, but sources of non-U.S. statistics will be covered as well.