Similarity and Matching of Neural Network Representations

Part of Advances in Neural Information Processing Systems 34 pre-proceedings (NeurIPS 2021)

Paper Supplemental

Bibtek download is not available in the pre-proceeding


Authors

Adrián Csiszárik, Péter Kőrösi-Szabó, Ákos Matszangosz, Gergely Papp, Dániel Varga

Abstract

We employ a toolset --- dubbed Dr. Frankenstein --- to analyse the similarity of representations in deep neural networks. With this toolset we aim to match the activations on given layers of two trained neural networks by joining them with a stitching layer. We demonstrate that the inner representations emerging in deep convolutional neural networks with the same architecture but different initialisations can be matched with a surprisingly high degree of accuracy even with a single, affine stitching layer. We choose the stitching layer from several possible classes of linear transformations and investigate their performance and properties. The task of matching representations is closely related to notions of similarity. Using this toolset we also provide a novel viewpoint on the current line of research regarding similarity indices of neural network representations: the perspective of the performance on a task.