Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385278446> ?p ?o ?g. }
- W4385278446 endingPage "722" @default.
- W4385278446 startingPage "697" @default.
- W4385278446 abstract "Abstract Modern neural network architectures are becoming larger and deeper, with increasing computational resources needed for training and inference. One approach toward handling this increased resource consumption is to use structured weight matrices. By exploiting structures in weight matrices, the computational complexity for propagating information through the network can be reduced. However, choosing the right structure is not trivial, especially since there are many different matrix structures and structure classes. In this paper, we give an overview over the four main matrix structure classes, namely semiseparable matrices, matrices of low displacement rank, hierarchical matrices and products of sparse matrices. We recapitulate the definitions of each structure class, present special structure subclasses, and provide references to research papers in which the structures are used in the domain of neural networks. We present two benchmarks comparing the classes. First, we benchmark the error for approximating different test matrices. Second, we compare the prediction performance of neural networks in which the weight matrix of the last layer is replaced by structured matrices. After presenting the benchmark results, we discuss open research questions related to the use of structured matrices in neural networks and highlight future research directions." @default.
- W4385278446 created "2023-07-27" @default.
- W4385278446 creator A5033656694 @default.
- W4385278446 creator A5074346065 @default.
- W4385278446 date "2023-07-26" @default.
- W4385278446 modified "2023-09-26" @default.
- W4385278446 title "Structured Matrices and Their Application in Neural Networks: A Survey" @default.
- W4385278446 cites W1506342804 @default.
- W4385278446 cites W1584657480 @default.
- W4385278446 cites W1588915715 @default.
- W4385278446 cites W1973786815 @default.
- W4385278446 cites W1983452621 @default.
- W4385278446 cites W1984439507 @default.
- W4385278446 cites W1987393840 @default.
- W4385278446 cites W1987724179 @default.
- W4385278446 cites W2001626771 @default.
- W4385278446 cites W2019166112 @default.
- W4385278446 cites W2026651112 @default.
- W4385278446 cites W2027828864 @default.
- W4385278446 cites W2038198231 @default.
- W4385278446 cites W2050746438 @default.
- W4385278446 cites W2051395078 @default.
- W4385278446 cites W2061171222 @default.
- W4385278446 cites W2062240003 @default.
- W4385278446 cites W2063961547 @default.
- W4385278446 cites W2083206954 @default.
- W4385278446 cites W2098602979 @default.
- W4385278446 cites W2110404184 @default.
- W4385278446 cites W2114977680 @default.
- W4385278446 cites W2132820941 @default.
- W4385278446 cites W2132984323 @default.
- W4385278446 cites W2145140212 @default.
- W4385278446 cites W2152332944 @default.
- W4385278446 cites W2157164240 @default.
- W4385278446 cites W21639112 @default.
- W4385278446 cites W2171506994 @default.
- W4385278446 cites W2188296733 @default.
- W4385278446 cites W2194775991 @default.
- W4385278446 cites W2245896225 @default.
- W4385278446 cites W2274086169 @default.
- W4385278446 cites W2293747114 @default.
- W4385278446 cites W2298381282 @default.
- W4385278446 cites W2558026684 @default.
- W4385278446 cites W2590462597 @default.
- W4385278446 cites W2593110912 @default.
- W4385278446 cites W2604319603 @default.
- W4385278446 cites W2741427986 @default.
- W4385278446 cites W2749641708 @default.
- W4385278446 cites W2765315405 @default.
- W4385278446 cites W2766447205 @default.
- W4385278446 cites W2766697789 @default.
- W4385278446 cites W2894977338 @default.
- W4385278446 cites W2905100748 @default.
- W4385278446 cites W2963230471 @default.
- W4385278446 cites W2963585199 @default.
- W4385278446 cites W2963809228 @default.
- W4385278446 cites W2964156591 @default.
- W4385278446 cites W2989600087 @default.
- W4385278446 cites W3037399553 @default.
- W4385278446 cites W3100023701 @default.
- W4385278446 cites W3127887098 @default.
- W4385278446 cites W3134229947 @default.
- W4385278446 cites W3154671241 @default.
- W4385278446 cites W4297811661 @default.
- W4385278446 cites W4297811735 @default.
- W4385278446 cites W4297838910 @default.
- W4385278446 cites W4312396463 @default.
- W4385278446 cites W608773355 @default.
- W4385278446 doi "https://doi.org/10.1007/s00354-023-00226-1" @default.
- W4385278446 hasPublicationYear "2023" @default.
- W4385278446 type Work @default.
- W4385278446 citedByCount "0" @default.
- W4385278446 crossrefType "journal-article" @default.
- W4385278446 hasAuthorship W4385278446A5033656694 @default.
- W4385278446 hasAuthorship W4385278446A5074346065 @default.
- W4385278446 hasBestOaLocation W43852784461 @default.
- W4385278446 hasConcept C106487976 @default.
- W4385278446 hasConcept C11413529 @default.
- W4385278446 hasConcept C114614502 @default.
- W4385278446 hasConcept C13280743 @default.
- W4385278446 hasConcept C134306372 @default.
- W4385278446 hasConcept C154945302 @default.
- W4385278446 hasConcept C159985019 @default.
- W4385278446 hasConcept C164226766 @default.
- W4385278446 hasConcept C185798385 @default.
- W4385278446 hasConcept C192562407 @default.
- W4385278446 hasConcept C205649164 @default.
- W4385278446 hasConcept C2776214188 @default.
- W4385278446 hasConcept C2777212361 @default.
- W4385278446 hasConcept C33923547 @default.
- W4385278446 hasConcept C36503486 @default.
- W4385278446 hasConcept C41008148 @default.
- W4385278446 hasConcept C50644808 @default.
- W4385278446 hasConcept C80444323 @default.
- W4385278446 hasConceptScore W4385278446C106487976 @default.
- W4385278446 hasConceptScore W4385278446C11413529 @default.
- W4385278446 hasConceptScore W4385278446C114614502 @default.
- W4385278446 hasConceptScore W4385278446C13280743 @default.