An explanation method for Siamese neural networks

TL;DR

This paper introduces a novel explanation method for Siamese neural networks that compares feature vectors to class prototypes and uses a specialized autoencoder for input reconstruction, demonstrated on MNIST.

Contribution

It presents a new explanation approach combining prototype comparison and autoencoder reconstruction tailored for Siamese neural networks.

Findings

Effective in identifying important features at the embedding level

Able to reconstruct inputs with relevant feature changes

Validated on MNIST dataset

Abstract

A new method for explaining the Siamese neural network is proposed. It uses the following main ideas. First, the explained feature vector is compared with the prototype of the corresponding class computed at the embedding level (the Siamese neural network output). The important features at this level are determined as features which are close to the same features of the prototype. Second, an autoencoder is trained in a special way in order to take into account the embedding level of the Si-amese network, and its decoder part is used for reconstructing input data with the corresponding changes. Numerical experiments with the well-known dataset MNIST illustrate the propose method.