Nth Absolute Root Mean Error

TL;DR

The paper introduces NARME, a novel loss function that significantly accelerates neural network training for regression tasks, reducing epochs needed and maintaining high accuracy, especially beneficial for large datasets.

Contribution

It proposes NARME, a new loss function that improves training speed and efficiency for neural network regression models compared to existing loss functions.

Findings

NARME reduces training epochs to about one-tenth of traditional loss functions.

NARME achieves high accuracy with less training time.

Experiments demonstrate faster convergence using NARME.

Abstract

Neural network training process takes long time when the size of training data is huge, without the large set of training values the neural network is unable to learn features. This dilemma between time and size of data is often solved using fast GPUs, but we present a better solution for a subset of those problems. To reduce the time for training a regression model using neural network we introduce a loss function called Nth Absolute Root Mean Error (NARME). It helps to train regression models much faster compared to other existing loss functions. Experiments show that in most use cases NARME reduces the required number of epochs to almost one-tenth of that required by other commonly used loss functions, and also achieves great accuracy in the small amount of time in which it was trained.