Matches in SemOpenAlex for { <https://semopenalex.org/work/W4310629408> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W4310629408 abstract "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 transformations defined by the model architecture, resulting in high-dimensional, difficult-to-interpret internal representations of input data. As DNNs become more ubiquitous across multiple sectors of our society, there is increasing recognition that mathematical methods are needed to aid analysts, researchers, and practitioners in understanding and interpreting how these models' internal representations relate to the final classification. In this paper, we apply cutting edge techniques from TDA with the goal of gaining insight into the interpretability of convolutional neural networks used for image classification. We use two common TDA approaches to explore several methods for modeling hidden-layer activations as high-dimensional point clouds, and provide experimental evidence that these point clouds capture valuable structural information about the model's process. First, we demonstrate that a distance metric based on persistent homology can be used to quantify meaningful differences between layers, and we discuss these distances in the broader context of existing representational similarity metrics for neural network interpretability. Second, we show that a mapper graph can provide semantic insight into how these models organize hierarchical class knowledge at each layer. These observations demonstrate that TDA is a useful tool to help deep learning practitioners unlock the hidden structures of their models." @default.
- W4310629408 created "2022-12-13" @default.
- W4310629408 creator A5016129197 @default.
- W4310629408 creator A5019903468 @default.
- W4310629408 creator A5020383361 @default.
- W4310629408 creator A5022505886 @default.
- W4310629408 creator A5031822529 @default.
- W4310629408 creator A5032787473 @default.
- W4310629408 creator A5047290243 @default.
- W4310629408 creator A5051236474 @default.
- W4310629408 creator A5059387152 @default.
- W4310629408 date "2022-11-30" @default.
- W4310629408 modified "2023-10-03" @default.
- W4310629408 title "Experimental Observations of the Topology of Convolutional Neural Network Activations" @default.
- W4310629408 doi "https://doi.org/10.48550/arxiv.2212.00222" @default.
- W4310629408 hasPublicationYear "2022" @default.
- W4310629408 type Work @default.
- W4310629408 citedByCount "0" @default.
- W4310629408 crossrefType "posted-content" @default.
- W4310629408 hasAuthorship W4310629408A5016129197 @default.
- W4310629408 hasAuthorship W4310629408A5019903468 @default.
- W4310629408 hasAuthorship W4310629408A5020383361 @default.
- W4310629408 hasAuthorship W4310629408A5022505886 @default.
- W4310629408 hasAuthorship W4310629408A5031822529 @default.
- W4310629408 hasAuthorship W4310629408A5032787473 @default.
- W4310629408 hasAuthorship W4310629408A5047290243 @default.
- W4310629408 hasAuthorship W4310629408A5051236474 @default.
- W4310629408 hasAuthorship W4310629408A5059387152 @default.
- W4310629408 hasBestOaLocation W43106294081 @default.
- W4310629408 hasConcept C108583219 @default.
- W4310629408 hasConcept C111919701 @default.
- W4310629408 hasConcept C11413529 @default.
- W4310629408 hasConcept C119857082 @default.
- W4310629408 hasConcept C138885662 @default.
- W4310629408 hasConcept C151730666 @default.
- W4310629408 hasConcept C154945302 @default.
- W4310629408 hasConcept C174348530 @default.
- W4310629408 hasConcept C199845137 @default.
- W4310629408 hasConcept C2776477805 @default.
- W4310629408 hasConcept C2777742833 @default.
- W4310629408 hasConcept C2779343474 @default.
- W4310629408 hasConcept C2781067378 @default.
- W4310629408 hasConcept C31258907 @default.
- W4310629408 hasConcept C41008148 @default.
- W4310629408 hasConcept C41895202 @default.
- W4310629408 hasConcept C50644808 @default.
- W4310629408 hasConcept C80444323 @default.
- W4310629408 hasConcept C81363708 @default.
- W4310629408 hasConcept C86803240 @default.
- W4310629408 hasConceptScore W4310629408C108583219 @default.
- W4310629408 hasConceptScore W4310629408C111919701 @default.
- W4310629408 hasConceptScore W4310629408C11413529 @default.
- W4310629408 hasConceptScore W4310629408C119857082 @default.
- W4310629408 hasConceptScore W4310629408C138885662 @default.
- W4310629408 hasConceptScore W4310629408C151730666 @default.
- W4310629408 hasConceptScore W4310629408C154945302 @default.
- W4310629408 hasConceptScore W4310629408C174348530 @default.
- W4310629408 hasConceptScore W4310629408C199845137 @default.
- W4310629408 hasConceptScore W4310629408C2776477805 @default.
- W4310629408 hasConceptScore W4310629408C2777742833 @default.
- W4310629408 hasConceptScore W4310629408C2779343474 @default.
- W4310629408 hasConceptScore W4310629408C2781067378 @default.
- W4310629408 hasConceptScore W4310629408C31258907 @default.
- W4310629408 hasConceptScore W4310629408C41008148 @default.
- W4310629408 hasConceptScore W4310629408C41895202 @default.
- W4310629408 hasConceptScore W4310629408C50644808 @default.
- W4310629408 hasConceptScore W4310629408C80444323 @default.
- W4310629408 hasConceptScore W4310629408C81363708 @default.
- W4310629408 hasConceptScore W4310629408C86803240 @default.
- W4310629408 hasLocation W43106294081 @default.
- W4310629408 hasLocation W43106294082 @default.
- W4310629408 hasOpenAccess W4310629408 @default.
- W4310629408 hasPrimaryLocation W43106294081 @default.
- W4310629408 hasRelatedWork W2731899572 @default.
- W4310629408 hasRelatedWork W3006943036 @default.
- W4310629408 hasRelatedWork W3129898729 @default.
- W4310629408 hasRelatedWork W3133861977 @default.
- W4310629408 hasRelatedWork W4200173597 @default.
- W4310629408 hasRelatedWork W4299487748 @default.
- W4310629408 hasRelatedWork W4310880831 @default.
- W4310629408 hasRelatedWork W4312417841 @default.
- W4310629408 hasRelatedWork W4321369474 @default.
- W4310629408 hasRelatedWork W4385957992 @default.
- W4310629408 isParatext "false" @default.
- W4310629408 isRetracted "false" @default.
- W4310629408 workType "article" @default.