The University of Arizona
Please note that this event has ended!

Interpretable & Tractable Machine Learning for Natural and Engineering Sciences.


Interpretable & Tractable Machine Learning for Natural and Engineering Sciences.
Series: TRIPODS Seminar
Location: ENR2 S210
Presenter: Michael Chertkov, Chair of the Program in Applied Mathematics at the University of Arizona.
Thanks to IT industry push, Machine Learning (ML) capabilities are in a phase of tremendous growth, and there is great opportunity to point these practically powerful tools toward modeling specific to applications, e.g. in natural and engineering sciences. The challenge is to incorporate domain expertise from traditional physical and engineering discipline scientific discovery into next-generation ML models. We propose to develop novel applied & theoretical mathematics and statistics, computational  and algorithmic, that extends cutting-edge ML tools and merge them with application-specific
knowledge stated in the form of constraints, symmetries, conservation laws, phenomenological assumptions and other examples of domain expertise regarding relevant degrees of freedom. The emerging Quantitative Machine Learning (QML) methodology bridges the two complementary poles -   application agnostic modern machine learning
(in particular deep learning), computationally efficient but lacking interpretability,  and science based tuning, highly interpretable but lacking automatization and implementation efficiency.  Different aspects of the QML methodology are illustrated on the following three enabling examples:
1.       Acceleration of Computational Fluid Mechanics Modeling and Simulations with Deep Learning [APS/DFD 2017-18, NIPS 2018 arXiv:1810.07785]
2.       Learning Graphical Models [Science  018/arXiv:1612.05024] and NIPS2016/arXiv:1605.07252]
3.       Gauges, Loops, and Polynomials for Partition Functions of Graphical Models [arXiv:1811.04713]


(Pizza, coffee & tea will be provided at 11:20am)