Explaining Bias In Your Data

Over the last five years, unfairness in machine learning has gone from almost unknown to hitting the headlines frequently, and new cases of unwanted bias introduced in automated processes are frequently discovered. However, there is still no "one-size-fits-all" standard machine learning tool to prevent and assess such bias. In this article, we will deal with how to explain unfairness in a machine learning algorithm. In 2014, in a report called Big Data: Seizing Opportunities and Preserving Values, the Executive Office of President Obama pointed out the fact that "big data technologies can cause societal harms beyond damages to privacy, such as discrimination against individuals and groups". This was the first time unfairness in machine learning had been officially recognized as a potential harm.