Focused Research Groups (FRG) grants are one of the most prestigious NSF awards. This year, one of the few awardees is the multi-university research initiative “FRG: Collaborative Research: Variationally Stable Neural Networks for Simulation, Learning, and Experimental Design of Complex Physical Systems” led by the Math Department at USC with PI Wolfgang Dahmen and Co-PIs Peter Binev, Wuchen Li, Hong Wang, and Zhu Wang. The other participating universities are UT Austin with PI Leszek Demkowicz, Portland State with PI Jay Gopalakrishnan, and Georgia Tech with PI Peng Chen.

The overarching goal of this project is to develop a mathematically rigorous framework for “learning” the underlying complex models from the given sources of information. The most stunning success of modern machine learning, especially deep learning, concerns so far mostly error tolerant applications. The particular ambition of the researchers working on this project is to significantly advance prediction capabilities through rigorous accuracy quantification and certification, as an indispensable feature of next generation simulation tools in science and technology.

 

• 

  Wolfgang Dahmen

  Principle Investigator (PI)

• 

  Peter Binev

  Co-PI

• 

  Wuchen Li

  Co-PI

• 

  Hong Wang

  Co-PI

• 

  Zhu Wang

  Co-PI