FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312286025> ?p ?o ?g. }

• W4312286025 abstract "Hyperbolic neural networks have been popular in the recent past due to their ability to represent hierarchical data sets effectively and efficiently. The challenge in developing these networks lies in the nonlinearity of the embedding space namely, the Hyperbolic space. Hyperbolic space is a homogeneous Riemannian manifold of the Lorentz group which is a semi-Riemannian manifold, i.e. a mani-fold equipped with an indefinite metric. Most existing methods (with some exceptions) use local linearization to define a variety of operations paralleling those used in traditional deep neural networks in Euclidean spaces. In this paper, we present a novel fully hyperbolic neural network which uses the concept of projections (embeddings) followed by an intrinsic aggregation and a nonlinearity all within the hyperbolic space. The novelty here lies in the projection which is designed to project data on to a lower-dimensional embedded hyperbolic space and hence leads to a nested hyperbolic space representation independently useful for dimensionality reduction. The main theoretical contribution is that the proposed embedding is proved to be isometric and equivariant under the Lorentz transformations, which are the natural isometric transformations in hyperbolic spaces. This projection is computationally efficient since it can be expressed by simple linear operations, and, due to the aforementioned equivariance property, it allows for weight sharing. The nested hyperbolic space representation is the core component of our network and therefore, we first compare this representation - independent of the network - with other dimensionality reduction methods such as tangent PCA, principal geodesic analysis (PGA) and HoroPCA. Based on this equivariant embedding, we develop a novel fully hyperbolic graph convolutional neural network architecture to learn the parameters of the projection. Finally, we present experiments demonstrating comparative performance of our network on several publicly available data sets." @default.
• W4312286025 created "2023-01-04" @default.
• W4312286025 creator A5012020834 @default.
• W4312286025 creator A5016378470 @default.
• W4312286025 creator A5049679429 @default.
• W4312286025 date "2022-06-01" @default.
• W4312286025 modified "2023-09-27" @default.
• W4312286025 title "Nested Hyperbolic Spaces for Dimensionality Reduction and Hyperbolic NN Design" @default.
• W4312286025 cites W1540708136 @default.
• W4312286025 cites W1907953363 @default.
• W4312286025 cites W1966521860 @default.
• W4312286025 cites W2045512849 @default.
• W4312286025 cites W2125949583 @default.
• W4312286025 cites W2153959628 @default.
• W4312286025 cites W2199357291 @default.
• W4312286025 cites W2295124130 @default.
• W4312286025 cites W2296883217 @default.
• W4312286025 cites W2777752764 @default.
• W4312286025 cites W2797520557 @default.
• W4312286025 cites W2962772276 @default.
• W4312286025 cites W2963480507 @default.
• W4312286025 cites W2982490267 @default.
• W4312286025 cites W3035102141 @default.
• W4312286025 cites W3164410012 @default.
• W4312286025 cites W3173288991 @default.
• W4312286025 cites W3177377177 @default.
• W4312286025 cites W4251002338 @default.
• W4312286025 doi "https://doi.org/10.1109/cvpr52688.2022.00045" @default.
• W4312286025 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/36911245" @default.
• W4312286025 hasPublicationYear "2022" @default.
• W4312286025 type Work @default.
• W4312286025 citedByCount "0" @default.
• W4312286025 crossrefType "proceedings-article" @default.
• W4312286025 hasAuthorship W4312286025A5012020834 @default.
• W4312286025 hasAuthorship W4312286025A5016378470 @default.
• W4312286025 hasAuthorship W4312286025A5049679429 @default.
• W4312286025 hasBestOaLocation W43122860252 @default.
• W4312286025 hasConcept C114614502 @default.
• W4312286025 hasConcept C134306372 @default.
• W4312286025 hasConcept C154945302 @default.
• W4312286025 hasConcept C157157409 @default.
• W4312286025 hasConcept C165818556 @default.
• W4312286025 hasConcept C167150821 @default.
• W4312286025 hasConcept C171036898 @default.
• W4312286025 hasConcept C184720557 @default.
• W4312286025 hasConcept C202444582 @default.
• W4312286025 hasConcept C33923547 @default.
• W4312286025 hasConcept C41008148 @default.
• W4312286025 hasConcept C41608201 @default.
• W4312286025 hasConcept C70518039 @default.
• W4312286025 hasConcept C70984080 @default.
• W4312286025 hasConcept C83677898 @default.
• W4312286025 hasConcept C92047909 @default.
• W4312286025 hasConcept C97654310 @default.
• W4312286025 hasConceptScore W4312286025C114614502 @default.
• W4312286025 hasConceptScore W4312286025C134306372 @default.
• W4312286025 hasConceptScore W4312286025C154945302 @default.
• W4312286025 hasConceptScore W4312286025C157157409 @default.
• W4312286025 hasConceptScore W4312286025C165818556 @default.
• W4312286025 hasConceptScore W4312286025C167150821 @default.
• W4312286025 hasConceptScore W4312286025C171036898 @default.
• W4312286025 hasConceptScore W4312286025C184720557 @default.
• W4312286025 hasConceptScore W4312286025C202444582 @default.
• W4312286025 hasConceptScore W4312286025C33923547 @default.
• W4312286025 hasConceptScore W4312286025C41008148 @default.
• W4312286025 hasConceptScore W4312286025C41608201 @default.
• W4312286025 hasConceptScore W4312286025C70518039 @default.
• W4312286025 hasConceptScore W4312286025C70984080 @default.
• W4312286025 hasConceptScore W4312286025C83677898 @default.
• W4312286025 hasConceptScore W4312286025C92047909 @default.
• W4312286025 hasConceptScore W4312286025C97654310 @default.
• W4312286025 hasFunder F4320306076 @default.
• W4312286025 hasLocation W43122860251 @default.
• W4312286025 hasLocation W43122860252 @default.
• W4312286025 hasLocation W43122860253 @default.
• W4312286025 hasLocation W43122860254 @default.
• W4312286025 hasOpenAccess W4312286025 @default.
• W4312286025 hasPrimaryLocation W43122860251 @default.
• W4312286025 hasRelatedWork W1541587918 @default.
• W4312286025 hasRelatedWork W1747263064 @default.
• W4312286025 hasRelatedWork W2001094874 @default.
• W4312286025 hasRelatedWork W2023866887 @default.
• W4312286025 hasRelatedWork W2045336066 @default.
• W4312286025 hasRelatedWork W2062999920 @default.
• W4312286025 hasRelatedWork W2950238559 @default.
• W4312286025 hasRelatedWork W2964208041 @default.
• W4312286025 hasRelatedWork W4297725804 @default.
• W4312286025 hasRelatedWork W4307680715 @default.
• W4312286025 isParatext "false" @default.
• W4312286025 isRetracted "false" @default.
• W4312286025 workType "article" @default.