Loading [a11y]/accessibility-menu.js
Recurrent Neural Network-Assisted Adaptive Sampling for Approximate Computing | IEEE Conference Publication | IEEE Xplore

Recurrent Neural Network-Assisted Adaptive Sampling for Approximate Computing

Publisher: IEEE

Abstract:

We propose an adaptive signal sampling approach that dynamically adjusts the sampling rate to approximate the local Nyquist rate of the signal. The proposed adaptive samp...View more

Abstract:

We propose an adaptive signal sampling approach that dynamically adjusts the sampling rate to approximate the local Nyquist rate of the signal. The proposed adaptive sampling approach consists of a recurrent neural network-based change detector that detects the point of frequency change and a local Nyquist rate estimator based on a multi-rate signal processing scheme. We empirically demonstrate that our adaptive sampling approach significantly reduces the overall sampling rate for various types of signals and therefore improves the computational efficiency of subsequent signal processing.
Date of Conference: 09-12 December 2019
Date Added to IEEE Xplore: 24 February 2020
ISBN Information:
Publisher: IEEE
Conference Location: Los Angeles, CA, USA

References

References is not available for this document.