English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Topological data analysis (TDA) is a branch of computational mathematics,bridging algebraic topology and data science, that provides compact,noise-robust representations of complex structures. Deep neural networks (DNNs)learn millions of parameters associated with a series of transformationsdefined by the model architecture, resulting in high-dimensional,difficult-to-interpret internal representations of input data. As DNNs becomemore ubiquitous across multiple sectors of our society, there is increasingrecognition that mathematical methods are needed to aid analysts, researchers,and practitioners in understanding and interpreting how these models' internalrepresentations relate to the final classification. In this paper, we applycutting edge techniques from TDA with the goal of gaining insight into theinterpretability of convolutional neural networks used for imageclassification. We use two common TDA approaches to explore several methods formodeling hidden-layer activations as high-dimensional point clouds, and provideexperimental evidence that these point clouds capture valuable structuralinformation about the model's process. First, we demonstrate that a distancemetric based on persistent homology can be used to quantify meaningfuldifferences between layers, and we discuss these distances in the broadercontext of existing representational similarity metrics for neural networkinterpretability. Second, we show that a mapper graph can provide semanticinsight into how these models organize hierarchical class knowledge at eachlayer. These observations demonstrate that TDA is a useful tool to help deeplearning practitioners unlock the hidden structures of their models.

Experimental Observations of the Topology of Convolutional Neural Network Activations