FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2947024452> ?p ?o ?g. }

• W2947024452 abstract "Datasets such as images, text, or movies are embedded in high-dimensional spaces. However, in important cases such as images of objects, the statistical structure in the data constrains samples to a manifold of dramatically lower dimensionality. Learning to identify and extract task-relevant variables from this embedded manifold is crucial when dealing with high-dimensional problems. We find that neural networks are often very effective at solving this task and investigate why. To this end, we apply state-of-the-art techniques for intrinsic dimensionality estimation to show that neural networks learn low-dimensional manifolds in two phases: first, dimensionality expansion driven by feature generation in initial layers, and second, dimensionality compression driven by the selection of task-relevant features in later layers. We model noise generated by Stochastic Gradient Descent and show how this noise balances the dimensionality of neural representations by inducing an effective regularization term in the loss. We highlight the important relationship between low-dimensional compressed representations and generalization properties of the network. Our work contributes by shedding light on the success of deep neural networks in disentangling data in high-dimensional space while achieving good generalization. Furthermore, it invites new learning strategies focused on optimizing measurable geometric properties of learned representations, beginning with their intrinsic dimensionality." @default.
• W2947024452 created "2019-06-07" @default.
• W2947024452 creator A5032055668 @default.
• W2947024452 creator A5042904819 @default.
• W2947024452 creator A5043037494 @default.
• W2947024452 creator A5046602924 @default.
• W2947024452 creator A5056428592 @default.
• W2947024452 creator A5090387403 @default.
• W2947024452 date "2019-06-02" @default.
• W2947024452 modified "2023-10-03" @default.
• W2947024452 title "Dimensionality compression and expansion in Deep Neural Networks." @default.
• W2947024452 cites W1496925227 @default.
• W2947024452 cites W1519654970 @default.
• W2947024452 cites W1528062093 @default.
• W2947024452 cites W1580448704 @default.
• W2947024452 cites W1592993660 @default.
• W2947024452 cites W2001141328 @default.
• W2947024452 cites W2017539895 @default.
• W2947024452 cites W2025768430 @default.
• W2947024452 cites W2029401646 @default.
• W2947024452 cites W2033708181 @default.
• W2947024452 cites W2083493892 @default.
• W2947024452 cites W2100495367 @default.
• W2947024452 cites W2105464873 @default.
• W2947024452 cites W2119479037 @default.
• W2947024452 cites W2152722485 @default.
• W2947024452 cites W2156287497 @default.
• W2947024452 cites W2157169955 @default.
• W2947024452 cites W2161278885 @default.
• W2947024452 cites W2163922914 @default.
• W2947024452 cites W2170567676 @default.
• W2947024452 cites W2255581062 @default.
• W2947024452 cites W2274287116 @default.
• W2947024452 cites W2467203645 @default.
• W2947024452 cites W2469085137 @default.
• W2947024452 cites W2513463761 @default.
• W2947024452 cites W2731468224 @default.
• W2947024452 cites W2763894180 @default.
• W2947024452 cites W2767382501 @default.
• W2947024452 cites W2767581643 @default.
• W2947024452 cites W2782762890 @default.
• W2947024452 cites W2786622092 @default.
• W2947024452 cites W2894573160 @default.
• W2947024452 cites W2903711666 @default.
• W2947024452 cites W2917507868 @default.
• W2947024452 cites W2962885042 @default.
• W2947024452 cites W2963069632 @default.
• W2947024452 cites W2963693826 @default.
• W2947024452 cites W2964231450 @default.
• W2947024452 cites W2964299589 @default.
• W2947024452 cites W2985958162 @default.
• W2947024452 cites W3004639598 @default.
• W2947024452 cites W3006348122 @default.
• W2947024452 cites W3021886971 @default.
• W2947024452 cites W3137695714 @default.
• W2947024452 hasPublicationYear "2019" @default.
• W2947024452 type Work @default.
• W2947024452 sameAs 2947024452 @default.
• W2947024452 citedByCount "7" @default.
• W2947024452 countsByYear W29470244522020 @default.
• W2947024452 countsByYear W29470244522021 @default.
• W2947024452 crossrefType "posted-content" @default.
• W2947024452 hasAuthorship W2947024452A5032055668 @default.
• W2947024452 hasAuthorship W2947024452A5042904819 @default.
• W2947024452 hasAuthorship W2947024452A5043037494 @default.
• W2947024452 hasAuthorship W2947024452A5046602924 @default.
• W2947024452 hasAuthorship W2947024452A5056428592 @default.
• W2947024452 hasAuthorship W2947024452A5090387403 @default.
• W2947024452 hasConcept C108583219 @default.
• W2947024452 hasConcept C111030470 @default.
• W2947024452 hasConcept C115961682 @default.
• W2947024452 hasConcept C119857082 @default.
• W2947024452 hasConcept C127413603 @default.
• W2947024452 hasConcept C134306372 @default.
• W2947024452 hasConcept C151876577 @default.
• W2947024452 hasConcept C153180895 @default.
• W2947024452 hasConcept C153258448 @default.
• W2947024452 hasConcept C154945302 @default.
• W2947024452 hasConcept C177148314 @default.
• W2947024452 hasConcept C206688291 @default.
• W2947024452 hasConcept C2776135515 @default.
• W2947024452 hasConcept C2984842247 @default.
• W2947024452 hasConcept C33923547 @default.
• W2947024452 hasConcept C41008148 @default.
• W2947024452 hasConcept C50644808 @default.
• W2947024452 hasConcept C529865628 @default.
• W2947024452 hasConcept C70518039 @default.
• W2947024452 hasConcept C78519656 @default.
• W2947024452 hasConcept C99498987 @default.
• W2947024452 hasConceptScore W2947024452C108583219 @default.
• W2947024452 hasConceptScore W2947024452C111030470 @default.
• W2947024452 hasConceptScore W2947024452C115961682 @default.
• W2947024452 hasConceptScore W2947024452C119857082 @default.
• W2947024452 hasConceptScore W2947024452C127413603 @default.
• W2947024452 hasConceptScore W2947024452C134306372 @default.
• W2947024452 hasConceptScore W2947024452C151876577 @default.
• W2947024452 hasConceptScore W2947024452C153180895 @default.
• W2947024452 hasConceptScore W2947024452C153258448 @default.
• W2947024452 hasConceptScore W2947024452C154945302 @default.
• W2947024452 hasConceptScore W2947024452C177148314 @default.

• next