Machine learning models are increasingly being used to make critical decisions that impact people’s lives. However, bias in training data, due to prejudice in labels and under- or oversampling, can result in models with unwanted bias. Discrimination can become an issue when machine learning models place certain privileged groups at systematic advantage and certain unprivileged groups at systematic disadvantage.

Ana Echeverri and Trisha Mahoney walk you through how to use the open source Python package AI Fairness 360, developed by IBM researchers. AI Fairness 360 is a comprehensive open source toolkit that empowers users with the metrics to check for unwanted bias in datasets and machine learning models and state-of-the-art algorithms to mitigate such bias.

You’ll learn which metric is most appropriate for a given use case, and when to use many of the different bias-mitigation algorithms provided in the toolkit. AI Fairness 360 provides an interactive experience as a gentle introduction to the concepts and capabilities of the toolkit for those unfamiliar with Python, as well as detailed tutorials for more advanced data scientists.

This event is sponsored by IBM.

Trisha Mahoney
Senior Technical Evangelist | IBM Machine Learning & AI

Trisha has spent the last 9 years working in high-tech firms doing product management/marketing roles in AI & Cloud (at IBM, Salesforce, Cisco and Smiths Group). Prior to that, Trisha spent 8 years working as a data scientist in the chemical detection space. She holds an Electrical Engineering degree and an MBA in Technology Management.