Machine Learning (statistical eng ineering) capabilities are in a phase of tremendous growth. Underlying the se advances is a strong and deep connection to various aspects of statisti cal physics. There is also a great opportunity in pointing these tools tow ard physical modeling. In this colloquium I \; illustrate the two-way flow of ideas between physics and statistical engineering on three example s from our team LANL. First\, I review the work on \; structure learni ng and statistical estimation in power system distribution (thus physical) networks. Then I describe recent progress in constructive understanding o f graph learning (on example of inverse Ising model) illustrating that the generic inverse task (of learning) is computationally easy in spite of th e fact that the direct problem (inference or sampling) is difficult. I con clude speculating how macro-scale models of physics (e.g. large eddy simul ations of turbulence) can be learned from micro-scale simulations (e.g. of Navier-Stocks equations).

\n\n\n DTSTART:20160304T203000Z LOCATION:Physics Building\, Room 204 SUMMARY:Physics Informed Machine Learning END:VEVENT END:VCALENDAR