In this talk we will look at machine learning, what it means for it to be robust and why that’s important. Existing and new methods under development will be discussed for making more robust machine learning available for the large messy data common in todays digital ecosystem.