# Sound source ranging using a feed-forward neural network with fitting-based early stopping

## TL;DR

This paper introduces FEAST, a fitting-based early stopping method for training feed-forward neural networks to improve sound source ranging accuracy in ocean waveguides, validated on simulated and real data.

## Contribution

The paper presents FEAST, a novel early stopping technique that evaluates range error without labeled test data, enhancing neural network ranging performance.

## Key findings

- FEAST effectively reduces range error during training.
- Improved ranging accuracy demonstrated on simulated data.
- Validated on experimental ocean waveguide data.

## Abstract

When a feed-forward neural network (FNN) is trained for source ranging in an ocean waveguide, it is difficult evaluating the range accuracy of the FNN on unlabeled test data. A fitting-based early stopping (FEAST) method is introduced to evaluate the range error of the FNN on test data where the distance of source is unknown. Based on FEAST, when the evaluated range error of the FNN reaches the minimum on test data, stopping training, which will help to improve the ranging accuracy of the FNN on the test data. The FEAST is demonstrated on simulated and experimental data.