An explanation method for Siamese neural networks
This addresses the need for interpretability in Siamese networks, but it is incremental as it builds on existing explanation techniques.
The paper tackles the problem of explaining Siamese neural networks by proposing a method that compares feature vectors with class prototypes and uses a specially trained autoencoder for reconstruction, with experiments conducted on the MNIST dataset.
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.