How Can Machine Learning Help Your Research Forward?

Machine learning is a buzzwords that conjures up visions of programming gurus and data magicians solving problems with little effort while others balk at the black-box nature and lack of first principles understanding. In this talk I hope to introduce some ways in which you can start to use powerful machine learning algorithms to solve certain classes of problems in ways that may be more generic than traditional approaches. I will use examples from a range of fields to demonstrate the power of machine learning, even though those field with access to large data sets have lead the charge. I will highlight differences between machine learning in physics and other data sciences. Finally, I will point out why a solid understanding of the underlying physical principles is a necessity to use machine learning in research with any success.