Paper Title
A Survey on Approximate Computing Techniques for Trainable Neural Network

Abstract
This study presents a review of the cutting-edge neural network's exploration of approximate computing. As most excellent among other energy-efficient computation methodologies, approximate computing has gotten significantly more research consideration over the past few years. This paper presents various approximate computing techniques used for quality management and increase energy efficiency in neural networks. We distinguish those techniques based on discrete key features to underscore their correlations and variations. This paper intends to give developers intuitions into approximate computing strategies and has encouraged many endeavors to make approximate computing a more common computing approach for future systems. Keywords - Approximate Computing, Neural Networks, Multitask Learning, Approximator.