A Low-Complexity Universal Scheme for Rate-Constrained Distributed Regression Using a Wireless Sensor Network

Download Now Free registration required

Executive Summary

The authors propose a scheme for rate-constrained distributed non-parametric regression using a wireless sensor network. The scheme is universal across a wide range of sensor noise models, including unbounded and non-additive noise; it has low complexity, requiring simple operations such as uniform scalar quantization with dither and message passing between neighboring nodes in the network; and attains mini-max optimality for regression functions in common smoothness classes. They present theoretical results on the trade-off between the compression rate, communication complexity of encoding, and the MSE and demonstrate empirical performance of the scheme using simulations.

  • Format: PDF
  • Size: 825.73 KB