Neural Network-Based Analog-to-Digital Converters

Abstract

In this chapter, we present an overview of the recent advances in analog-to-digital converter (ADC) neural networks. Biological neural networks consist of natural binarization reflected by the neurosynaptic processes. This natural analog-to-binary conversion ability of neurons can be modeled to emulate analog-to-digital conversion using a set of nonlinear circuit elements and existing artificial neural network models. Since one neuron during processing consumes on average only about half nanowatts of power, neurons can perform highly energy-efficient operations, including pattern recognition. Analog-to-digital conversion itself is an example of simple pattern recognition where input analog signal can be presented in one of the 2 N different patterns for N bits. The classical configuration of neural network-based ADC is Hopfield neural network ADC. Improved designs, such as modified Hopfield network ADC, T-model neural ADC, and multilevel neurons-based neural ADC, will be discussed. In addition, the latest architecture designs of neural ADC such as hybrid complementary metal-oxide semiconductor (CMOS)-memristor Hopfield ADC are covered at the end of this chapter.

A Domain-Theoretic Framework for Robustness Analysis of Neural Networks

“…As such, lipDT is suitable for robustness analysis of deeper networks that appear in safety-critical engineering applications with low dimensional input, e. g., in power electronics [DWB19], analog-to-digital converters [TJ18], and solution of partial differential equations (PDEs) [Rud13]. When no bisections are required, lipDT can be useful for analysis of larger networks as well, as shown in Experiment 2 on the MNIST dataset (Table 3).…”

An Overview of Neuromorphic Computing for Artificial Intelligence Enabled Hardware-Based Hopfield Neural Network

“…The main idea of the Hopfield is the use of the hardware circuit to simulate neural network optimization process. This process can be fast that takes an analog circuit processing advantage rather than digital circuit [59]. Unlike the software realization of the Hopfield neural network, the hardware implementation of the algorithm makes brain-like computations possible [60].…”