I joined CIRA in July 2019 to support its use of machine learning (ML) techniques. I am at CIRA half-time, and am a research faculty member in Electrical and Computer Engineering at Colorado State U. during my remaining time.

I work on selected high priority CIRA projects, while also providing general ML support for other CIRA projects, as well as teaching ML skills through introductory and advanced lessons for CIRA folks.

My key interests for CIRA activities include

1. physics-guided machine learning, i.e. methods for merging science knowledge into machine learning algorithms for better generalization and robustness;
2. leveraging visualization and attribution techniques for neural networks to gain more understanding of their internal reasoning for weather/climate applications;
3. utilizing self-supervised learning to make the most out of large data sets with few labels; and
4. building strong collaborations with colleagues in academia working on related topics.

I am NOT a computer scientist by training, but a mathematician/engineer who has used machine learning for several years for applications in climate science, in close collaboration with climate scientists. So don’t expect me to know everything on the spot about every single machine learning method out there, but do expect me to have a solid understanding of how to apply machine learning methods specifically to physical systems, and to stay informed of new methods that could be useful for this purpose.

I received the equivalent of B.S. and M.S. degrees in mathematics from the University of Karlsruhe, Germany (known these days as KIT), and M.S. and Ph.D. degrees in Mechanical Engineering from the Johns Hopkins University in Baltimore, MD.  I serve on the steering committee of the annual Climate Informatics (CI) workshop series and of the NSF-funded research coordination network on Intelligent Systems for the Geosciences (IS-GEO).

The best way to get in touch with me is to send email to iebert@colostate.edu.