Spotlight
in
Workshop: Topological, Algebraic, and Geometric Pattern Recognition with Applications Workshop Proposal
Hamming Similarity and Graph Laplacians for Class Partitioning and Adversarial Image Detection
Huma Jamil · Yajing Liu · Turgay Caglar · Christina Cole · Nathaniel Blanchard · Christopher Peterson · Michael Kirby
Researchers typically investigate neural network representations by examining activation outputs for one or more layers of a network. Here, we investigate the potential for ReLU activation patterns (encoded as bit vectors) to aid in understanding and interpreting the behavior of neural networks. We utilize Representational Dissimilarity Matrices (RDMs) to investigate the coherence of data within the embedding spaces of a deep neural network. From each layer of a network, we extract and utilize bit vectors to construct similarity scores between images. From these similarity scores, we build a similarity matrix for a collection of images drawn from 2 classes. We then apply Fiedler partitioning to the associated Laplacian matrix to separate the classes. Our results indicate, through bit vector representations, that the network continues to refine class detectability with the last ReLU layer achieving better than 95% separation accuracy. Additionally, we demonstrate that bit vectors aid in adversarial image detection, again achieving over 95% accuracy in separating adversarial and non-adversarial images using a simple classifier.