Adaptive Machine Learning Techniques for Signal Indentification, Classification, and Recovery

Abstract

The goal of this research is to harness the power of generative neural networks for the problem of compressed sensing, i.e. recovering unknown signals from noisy measurements. We have recently applied neural networks to this domain, performing significantly better than classical signal recovery methods.More recent approaches with neural networks are impractical in that they require a large training set of signals; thus not only are these approaches expensive, but they are also restricted to generating simple signals. In contrast, we require no training set and is consequently able to reconstruct arbitrarily complex signals. For these reasons, there is great potential for applying our signal recovery method to a wide array of problems. The proposed research will be demonstrating the power of our method and its potential for solving signal recovery problems in communications, signal processing, joint radar/communications and related problems.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 20, 2019
Source ID
N000141912590

Entities

People

  • Sriram Vishwanath

Organizations

  • Office of Naval Research
  • United States Navy
  • University of Texas at Austin

Tags

Fields of Study

  • Engineering

Readers

  • Distributed Systems and Data Platform Development
  • Neural Network Machine Learning.

Technology Areas

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