AI Fairness 360: Detect and remediate bias in Machines Learning Datasets and Models

Animesh Singh

Audience level:
Novice

Description

Because flaws and biases may not be easy to detect without the right tool, we have launched  AI Fairness 360, an open source library to help detect and remove bias in machine learning models and data sets. We will share lessons learned while using AI Fairness 360 and demonstrate how to leverage it to detect and de bias models during pre-processing, in-processing, and post-processing.

Abstract

At this pivotal point in our history, the values we’re passing on to the next generation are once again in the spotlight. But this time, we need to understand how to pass on the concept of fairness to the next generation of artificial intelligence (AI) systems. The application of AI algorithms in domains such as criminal justice, credit scoring, and hiring holds unlimited promise. At the same time, it raises legitimate concerns about algorithmic fairness — AI systems are deciding everything from which resumes are considered, to which insurance claims will be accepted, to who gets their  mortgage loan approved, and even to who receives a parole.

What we need is a “comprehensive bias pipeline” that fully integrates into the AI lifecycle. Such a pipleine requires a robust set of checkers, “de-biasing” algorithms, and bias explanations. Because flaws and biases may not be easy to detect without the right tool, we have launched  AI Fairness 360, an open source library to help detect and remove bias in machine learning models and data sets.

The AI Fairness 360 Python package includes a comprehensive set of metrics for data sets and models to test for biases, explanations for these metrics, and algorithms to mitigate bias in data sets and models. The research community worked together to create 30 fairness metrics and nine state-of-the-art bias mitigation algorithms. 

We will share lessons learned while using AI Fairness 360 and demonstrate how to leverage it to detect and de bias models during pre-processing, in-processing, and post-processing. We will explain how to take these practices and apply them on training in on a more robust environment using Fabric for Deep Learning (FfDL, pronounced “fiddle”) which provides a consistent way to run various scalable deep learning frameworks as a service on Kubernetes.

Subscribe to Receive PyData Updates

Tickets

Get Now