Novel Computational Harmonic Analysis Techniques for Machine Learning and Inverse Problems

Abstract

This program aims to develop a completely new paradigm for machine learning (ML), and deep learning, basedmore solidly on principles of computational harmonic analysis (CHA) and approximation theory, which circumvents several recently discovered drawbacks of traditional ML (in particular, VC theory). The resulting theory will be validated with a wide class of regression, classification, and inverse problems problems in ML with applications of importance to the DoD and DoN mission, e.g., imaging and automatic target recognition (ATR) in electro-optical (EO) and Synthetic Aperture Radar (SAR) / Inverse SAR (ISAR) data.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 11, 2023
Source ID
N000142312790

Entities

People

  • Hrushikesh Mhaskar

Organizations

  • Claremont Graduate University
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Calculus or Mathematical Analysis
  • Distributed Systems and Data Platform Development
  • Radar Systems Engineering.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks