This is my work for the ML Lunch & Learn at Sailpoint. The directory contains the demo code + slides.
Title: LIME is the new black! How recent advances open up ML black-box models and help explain AI decisions.
Keeping up with all the hype surrounding Machine Learning and Deep Learning, these technologies are usually perceived utilizing black-box models because they are so complex for human brains to comprehend. Nevertheless, ML and DL are becoming more popular in advanced applications like autonomous driving, language translation, speech recognition, etc. The question is: can you trust these models in making the decisions for you?In this talk, we will introduce a few methods attempting to explain ML & DL models either globally or locally for a particular decision. Focusing on LIME, and SHAP, we will talk more about how to choose the best method given the problems we are trying to solve.