Matches in SemOpenAlex for { <https://semopenalex.org/work/W4361020224> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4361020224 abstract "Distributed sparse block codes (SBCs) exhibit compact representations for encoding and manipulating symbolic data structures using fixed-with vectors. One major challenge however is to disentangle, or factorize, such data structures into their constituent elements without having to search through all possible combinations. This factorization becomes more challenging when queried by noisy SBCs wherein symbol representations are relaxed due to perceptual uncertainty and approximations made when modern neural networks are used to generate the query vectors. To address these challenges, we first propose a fast and highly accurate method for factorizing a more flexible and hence generalized form of SBCs, dubbed GSBCs. Our iterative factorizer introduces a threshold-based nonlinear activation, a conditional random sampling, and an $ell_infty$-based similarity metric. Its random sampling mechanism in combination with the search in superposition allows to analytically determine the expected number of decoding iterations, which matches the empirical observations up to the GSBC's bundling capacity. Secondly, the proposed factorizer maintains its high accuracy when queried by noisy product vectors generated using deep convolutional neural networks (CNNs). This facilitates its application in replacing the large fully connected layer (FCL) in CNNs, whereby C trainable class vectors, or attribute combinations, can be implicitly represented by our factorizer having F-factor codebooks, each with $sqrt[leftroot{-2}uproot{2}F]{C}$ fixed codevectors. We provide a methodology to flexibly integrate our factorizer in the classification layer of CNNs with a novel loss function. We demonstrate the feasibility of our method on four deep CNN architectures over CIFAR-100, ImageNet-1K, and RAVEN datasets. In all use cases, the number of parameters and operations are significantly reduced compared to the FCL." @default.
- W4361020224 created "2023-03-30" @default.
- W4361020224 creator A5008994084 @default.
- W4361020224 creator A5017236774 @default.
- W4361020224 creator A5026984722 @default.
- W4361020224 creator A5043163538 @default.
- W4361020224 creator A5043408422 @default.
- W4361020224 creator A5048112794 @default.
- W4361020224 creator A5048981968 @default.
- W4361020224 creator A5050167359 @default.
- W4361020224 creator A5068274229 @default.
- W4361020224 date "2023-03-24" @default.
- W4361020224 modified "2023-10-17" @default.
- W4361020224 title "Factorizers for Distributed Sparse Block Codes" @default.
- W4361020224 doi "https://doi.org/10.48550/arxiv.2303.13957" @default.
- W4361020224 hasPublicationYear "2023" @default.
- W4361020224 type Work @default.
- W4361020224 citedByCount "0" @default.
- W4361020224 crossrefType "posted-content" @default.
- W4361020224 hasAuthorship W4361020224A5008994084 @default.
- W4361020224 hasAuthorship W4361020224A5017236774 @default.
- W4361020224 hasAuthorship W4361020224A5026984722 @default.
- W4361020224 hasAuthorship W4361020224A5043163538 @default.
- W4361020224 hasAuthorship W4361020224A5043408422 @default.
- W4361020224 hasAuthorship W4361020224A5048112794 @default.
- W4361020224 hasAuthorship W4361020224A5048981968 @default.
- W4361020224 hasAuthorship W4361020224A5050167359 @default.
- W4361020224 hasAuthorship W4361020224A5068274229 @default.
- W4361020224 hasBestOaLocation W43610202241 @default.
- W4361020224 hasConcept C11413529 @default.
- W4361020224 hasConcept C114614502 @default.
- W4361020224 hasConcept C134306372 @default.
- W4361020224 hasConcept C153180895 @default.
- W4361020224 hasConcept C154945302 @default.
- W4361020224 hasConcept C162324750 @default.
- W4361020224 hasConcept C176217482 @default.
- W4361020224 hasConcept C187834632 @default.
- W4361020224 hasConcept C21547014 @default.
- W4361020224 hasConcept C27753989 @default.
- W4361020224 hasConcept C2777210771 @default.
- W4361020224 hasConcept C33923547 @default.
- W4361020224 hasConcept C41008148 @default.
- W4361020224 hasConcept C57273362 @default.
- W4361020224 hasConcept C80444323 @default.
- W4361020224 hasConcept C81363708 @default.
- W4361020224 hasConceptScore W4361020224C11413529 @default.
- W4361020224 hasConceptScore W4361020224C114614502 @default.
- W4361020224 hasConceptScore W4361020224C134306372 @default.
- W4361020224 hasConceptScore W4361020224C153180895 @default.
- W4361020224 hasConceptScore W4361020224C154945302 @default.
- W4361020224 hasConceptScore W4361020224C162324750 @default.
- W4361020224 hasConceptScore W4361020224C176217482 @default.
- W4361020224 hasConceptScore W4361020224C187834632 @default.
- W4361020224 hasConceptScore W4361020224C21547014 @default.
- W4361020224 hasConceptScore W4361020224C27753989 @default.
- W4361020224 hasConceptScore W4361020224C2777210771 @default.
- W4361020224 hasConceptScore W4361020224C33923547 @default.
- W4361020224 hasConceptScore W4361020224C41008148 @default.
- W4361020224 hasConceptScore W4361020224C57273362 @default.
- W4361020224 hasConceptScore W4361020224C80444323 @default.
- W4361020224 hasConceptScore W4361020224C81363708 @default.
- W4361020224 hasLocation W43610202241 @default.
- W4361020224 hasOpenAccess W4361020224 @default.
- W4361020224 hasPrimaryLocation W43610202241 @default.
- W4361020224 hasRelatedWork W1991750312 @default.
- W4361020224 hasRelatedWork W2026169332 @default.
- W4361020224 hasRelatedWork W2352784731 @default.
- W4361020224 hasRelatedWork W2358034992 @default.
- W4361020224 hasRelatedWork W2382616524 @default.
- W4361020224 hasRelatedWork W2735477435 @default.
- W4361020224 hasRelatedWork W2748454020 @default.
- W4361020224 hasRelatedWork W2998526951 @default.
- W4361020224 hasRelatedWork W3052481912 @default.
- W4361020224 hasRelatedWork W3119610945 @default.
- W4361020224 isParatext "false" @default.
- W4361020224 isRetracted "false" @default.
- W4361020224 workType "article" @default.