Prof. Feldman is a worldwide leader and main developer both in academy and industry for the research field known as core sets: a provably small and problem dependent data reduction.  While the theory is based on deep computational geometry, (current\modern) applications include Machine/Deep learning of Big Data, Robotics, Computer Vision, Cybersecurity, and many more. 


Turning Big Data Into Tiny Data: Constant-Size Coresets for -Means, PCA, and Projective Clustering
Dan Feldman, Melanie Schmidt, Christian Sohler; SIAM Journal on Computing 49 (3), 601-657
Real-time EEG classification via coresets for BCI applications
E Netzer, A Frid, D Feldman; Engineering Applications of Artificial Intelligence 89, 103455
Deterministic Coresets for k-Means of Big Sparse Data
A Barger, D Feldman; Algorithms 13 (4), 92