Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287121275> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4287121275 abstract "Motivated by the vast success of deep convolutional networks, there is a great interest in generalizing convolutions to non-Euclidean manifolds. A major complication in comparison to flat spaces is that it is unclear in which alignment a convolution kernel should be applied on a manifold. The underlying reason for this ambiguity is that general manifolds do not come with a canonical choice of reference frames (gauge). Kernels and features therefore have to be expressed relative to arbitrary coordinates. We argue that the particular choice of coordinatization should not affect a network's inference -- it should be coordinate independent. A simultaneous demand for coordinate independence and weight sharing is shown to result in a requirement on the network to be equivariant under local gauge transformations (changes of local reference frames). The ambiguity of reference frames depends thereby on the G-structure of the manifold, such that the necessary level of gauge equivariance is prescribed by the corresponding structure group G. Coordinate independent convolutions are proven to be equivariant w.r.t. those isometries that are symmetries of the G-structure. The resulting theory is formulated in a coordinate free fashion in terms of fiber bundles. To exemplify the design of coordinate independent convolutions, we implement a convolutional network on the Mobius strip. The generality of our differential geometric formulation of convolutional networks is demonstrated by an extensive literature review which explains a large number of Euclidean CNNs, spherical CNNs and CNNs on general surfaces as specific instances of coordinate independent convolutions." @default.
- W4287121275 created "2022-07-25" @default.
- W4287121275 creator A5005753034 @default.
- W4287121275 creator A5059876635 @default.
- W4287121275 creator A5087368991 @default.
- W4287121275 creator A5088933660 @default.
- W4287121275 date "2021-06-10" @default.
- W4287121275 modified "2023-09-30" @default.
- W4287121275 title "Coordinate Independent Convolutional Networks -- Isometry and Gauge Equivariant Convolutions on Riemannian Manifolds" @default.
- W4287121275 doi "https://doi.org/10.48550/arxiv.2106.06020" @default.
- W4287121275 hasPublicationYear "2021" @default.
- W4287121275 type Work @default.
- W4287121275 citedByCount "0" @default.
- W4287121275 crossrefType "posted-content" @default.
- W4287121275 hasAuthorship W4287121275A5005753034 @default.
- W4287121275 hasAuthorship W4287121275A5059876635 @default.
- W4287121275 hasAuthorship W4287121275A5087368991 @default.
- W4287121275 hasAuthorship W4287121275A5088933660 @default.
- W4287121275 hasBestOaLocation W42871212751 @default.
- W4287121275 hasConcept C114614502 @default.
- W4287121275 hasConcept C127413603 @default.
- W4287121275 hasConcept C146710177 @default.
- W4287121275 hasConcept C154945302 @default.
- W4287121275 hasConcept C159985019 @default.
- W4287121275 hasConcept C171036898 @default.
- W4287121275 hasConcept C184720557 @default.
- W4287121275 hasConcept C192562407 @default.
- W4287121275 hasConcept C202444582 @default.
- W4287121275 hasConcept C2778134712 @default.
- W4287121275 hasConcept C33923547 @default.
- W4287121275 hasConcept C41008148 @default.
- W4287121275 hasConcept C45347329 @default.
- W4287121275 hasConcept C50644808 @default.
- W4287121275 hasConcept C529865628 @default.
- W4287121275 hasConcept C56907000 @default.
- W4287121275 hasConcept C74193536 @default.
- W4287121275 hasConcept C78519656 @default.
- W4287121275 hasConcept C82457910 @default.
- W4287121275 hasConcept C95857938 @default.
- W4287121275 hasConceptScore W4287121275C114614502 @default.
- W4287121275 hasConceptScore W4287121275C127413603 @default.
- W4287121275 hasConceptScore W4287121275C146710177 @default.
- W4287121275 hasConceptScore W4287121275C154945302 @default.
- W4287121275 hasConceptScore W4287121275C159985019 @default.
- W4287121275 hasConceptScore W4287121275C171036898 @default.
- W4287121275 hasConceptScore W4287121275C184720557 @default.
- W4287121275 hasConceptScore W4287121275C192562407 @default.
- W4287121275 hasConceptScore W4287121275C202444582 @default.
- W4287121275 hasConceptScore W4287121275C2778134712 @default.
- W4287121275 hasConceptScore W4287121275C33923547 @default.
- W4287121275 hasConceptScore W4287121275C41008148 @default.
- W4287121275 hasConceptScore W4287121275C45347329 @default.
- W4287121275 hasConceptScore W4287121275C50644808 @default.
- W4287121275 hasConceptScore W4287121275C529865628 @default.
- W4287121275 hasConceptScore W4287121275C56907000 @default.
- W4287121275 hasConceptScore W4287121275C74193536 @default.
- W4287121275 hasConceptScore W4287121275C78519656 @default.
- W4287121275 hasConceptScore W4287121275C82457910 @default.
- W4287121275 hasConceptScore W4287121275C95857938 @default.
- W4287121275 hasLocation W42871212751 @default.
- W4287121275 hasOpenAccess W4287121275 @default.
- W4287121275 hasPrimaryLocation W42871212751 @default.
- W4287121275 hasRelatedWork W1671228931 @default.
- W4287121275 hasRelatedWork W1973842768 @default.
- W4287121275 hasRelatedWork W2089528141 @default.
- W4287121275 hasRelatedWork W2558695868 @default.
- W4287121275 hasRelatedWork W2963326435 @default.
- W4287121275 hasRelatedWork W2963854314 @default.
- W4287121275 hasRelatedWork W2964134901 @default.
- W4287121275 hasRelatedWork W3171974924 @default.
- W4287121275 hasRelatedWork W3194528063 @default.
- W4287121275 hasRelatedWork W4287121275 @default.
- W4287121275 isParatext "false" @default.
- W4287121275 isRetracted "false" @default.
- W4287121275 workType "article" @default.