A powerful principle is to quantify the group behavior of the points, through problem-specific structures, that in turn instruct low-complexity approximations.
Pablo BarcelóInstitute for Mathematical and Computational Engineering, PUC Chile & IMFD Chile & CENIA ChileVerified email at uc.cl. Follow. Mircea Petrache.
Assistant Professor, Pontificia Catolica Universidad de Chile. Email: mpetrache AT mat.uc.cl. Home-page: https://sites.google.com/site/mircpetrache/home ...
Mircea Petrache
>
Affiliation: Pontifical Catholic University of Chile
Research interests: Deep Learning, Point Configurations, Calculus of Variations, and more
I do mathematics research, recently interested in Data Geometry and its relation with Deep Learning. In the past years, I have been devepopping mathematical ...
I am interested in Calculus of Variations, PDEs, Geometric Measure Theory, with special emphasis on interactions involving Probability, Optimization and ...
Doing maths research (geom. analysis, metric geometry, large point configs., opt. transport, etc) since ~2010, deep learning and applied research since ~2020.
Promoting openness in scientific communication and the peer-review process.
Dissertation: Weak Bundle Structures and a Variational Approach to Yang-Mills Lagrangians in Supercritical Dimensions. Advisor 1: Tristan Rivière.
677 Followers, 1221 Following, 201 Posts - See Instagram photos and videos from Mircea Petrache (@petrachemircea)
We introduce and study the unconstrained polarization (or Chebyshev) problem which requires to find an N-point configuration that maximizes the minimum value of ...