The Need for Fair and Interpretable Models in Machine Learning

As machine learning systems are increasingly being integrated into various industries and everyday life, there is an increasing need to hold such systems accountable. Especially in domains like healthcare, finance and the criminal justice system if there is bias in the decisions made by machine leanring or AI powered systems then that have destating consequences. One way to mitigate such problems is via machine learning models which are interpretable and fair. This talk will elucidate the various notions of interpretability and explanations in AI and machine learning systems, and why they will become an integral feature of machine learning systems in the near future.