Machine Learning's Poor Fit for Real Data

There's a growing sentiment out there with all the wonderful things happening in artificial intelligence, machine learning, and data science that these technologies are ready to solve all the things (including how to kill all humans). The reality is there are still a bunch of significant hurdles between us and the AI dystopia/utopia. One big one that is the main impetus behind my research is the disconnect between the statistical foundations of machine learning and how real data works. Machine learning technology is built on a foundation of formal theory. Statistical ideas, computer science algorithms, and information-theoretic concepts integrate to yield practical methods that analyze large, noisy data sets to train actionable and predictive models.