Matches in SemOpenAlex for { <https://semopenalex.org/work/W3113017313> ?p ?o ?g. }
- W3113017313 abstract "Recently, deep convolutional neural networks (DCNNs) have attained human-level performances on challenging object recognition tasks owing to their complex internal representation. However, it remains unclear how objects are represented in DCNNs with an overwhelming number of features and non-linear operations. In parallel, the same question has been extensively studied in primates' brain, and three types of coding schemes have been found: one object is coded by the entire neuronal population (distributed coding), or by one single neuron (local coding), or by a subset of neuronal population (sparse coding). Here we asked whether DCNNs adopted any of these coding schemes to represent objects. Specifically, we used the population sparseness index, which is widely-used in neurophysiological studies on primates' brain, to characterize the degree of sparseness at each layer in representative DCNNs pretrained for object categorization. We found that the sparse coding scheme was adopted at all layers of the DCNNs, and the degree of sparseness increased along the hierarchy. That is, the coding scheme shifted from distributed-like coding at lower layers to local-like coding at higher layers. Further, the degree of sparseness was positively correlated with DCNNs' performance in object categorization, suggesting that the coding scheme was related to behavioral performance. Finally, with the lesion approach, we demonstrated that both external learning experiences and built-in gating operations were necessary to construct such a hierarchical coding scheme. In sum, our study provides direct evidence that DCNNs adopted a hierarchically-evolved sparse coding scheme as the biological brain does, suggesting the possibility of an implementation-independent principle underling object recognition." @default.
- W3113017313 created "2020-12-21" @default.
- W3113017313 creator A5038594329 @default.
- W3113017313 creator A5053707540 @default.
- W3113017313 creator A5057117256 @default.
- W3113017313 date "2020-12-09" @default.
- W3113017313 modified "2023-10-13" @default.
- W3113017313 title "Hierarchical Sparse Coding of Objects in Deep Convolutional Neural Networks" @default.
- W3113017313 cites W1677182931 @default.
- W3113017313 cites W1839558765 @default.
- W3113017313 cites W1892402995 @default.
- W3113017313 cites W1934410531 @default.
- W3113017313 cites W1980824976 @default.
- W3113017313 cites W1980880162 @default.
- W3113017313 cites W1993533483 @default.
- W3113017313 cites W2004900646 @default.
- W3113017313 cites W2013239224 @default.
- W3113017313 cites W2029070203 @default.
- W3113017313 cites W2049353653 @default.
- W3113017313 cites W2058670155 @default.
- W3113017313 cites W2060582698 @default.
- W3113017313 cites W2072016696 @default.
- W3113017313 cites W2081580037 @default.
- W3113017313 cites W2085927826 @default.
- W3113017313 cites W2086628957 @default.
- W3113017313 cites W2086664483 @default.
- W3113017313 cites W2091845343 @default.
- W3113017313 cites W2097117768 @default.
- W3113017313 cites W2117539524 @default.
- W3113017313 cites W2139927483 @default.
- W3113017313 cites W2145889472 @default.
- W3113017313 cites W2160921898 @default.
- W3113017313 cites W2160996222 @default.
- W3113017313 cites W2528057474 @default.
- W3113017313 cites W2582586363 @default.
- W3113017313 cites W2900442071 @default.
- W3113017313 cites W2919115771 @default.
- W3113017313 cites W2952760759 @default.
- W3113017313 cites W2954137290 @default.
- W3113017313 cites W2984114831 @default.
- W3113017313 cites W3108571598 @default.
- W3113017313 cites W4253947814 @default.
- W3113017313 doi "https://doi.org/10.3389/fncom.2020.578158" @default.
- W3113017313 hasPubMedCentralId "https://www.ncbi.nlm.nih.gov/pmc/articles/7755594" @default.
- W3113017313 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/33362499" @default.
- W3113017313 hasPublicationYear "2020" @default.
- W3113017313 type Work @default.
- W3113017313 sameAs 3113017313 @default.
- W3113017313 citedByCount "5" @default.
- W3113017313 countsByYear W31130173132021 @default.
- W3113017313 countsByYear W31130173132022 @default.
- W3113017313 countsByYear W31130173132023 @default.
- W3113017313 crossrefType "journal-article" @default.
- W3113017313 hasAuthorship W3113017313A5038594329 @default.
- W3113017313 hasAuthorship W3113017313A5053707540 @default.
- W3113017313 hasAuthorship W3113017313A5057117256 @default.
- W3113017313 hasBestOaLocation W31130173131 @default.
- W3113017313 hasConcept C105795698 @default.
- W3113017313 hasConcept C144024400 @default.
- W3113017313 hasConcept C149923435 @default.
- W3113017313 hasConcept C153180895 @default.
- W3113017313 hasConcept C154945302 @default.
- W3113017313 hasConcept C179518139 @default.
- W3113017313 hasConcept C2908647359 @default.
- W3113017313 hasConcept C33923547 @default.
- W3113017313 hasConcept C41008148 @default.
- W3113017313 hasConcept C77637269 @default.
- W3113017313 hasConcept C81363708 @default.
- W3113017313 hasConcept C94124525 @default.
- W3113017313 hasConceptScore W3113017313C105795698 @default.
- W3113017313 hasConceptScore W3113017313C144024400 @default.
- W3113017313 hasConceptScore W3113017313C149923435 @default.
- W3113017313 hasConceptScore W3113017313C153180895 @default.
- W3113017313 hasConceptScore W3113017313C154945302 @default.
- W3113017313 hasConceptScore W3113017313C179518139 @default.
- W3113017313 hasConceptScore W3113017313C2908647359 @default.
- W3113017313 hasConceptScore W3113017313C33923547 @default.
- W3113017313 hasConceptScore W3113017313C41008148 @default.
- W3113017313 hasConceptScore W3113017313C77637269 @default.
- W3113017313 hasConceptScore W3113017313C81363708 @default.
- W3113017313 hasConceptScore W3113017313C94124525 @default.
- W3113017313 hasFunder F4320321001 @default.
- W3113017313 hasFunder F4320335777 @default.
- W3113017313 hasLocation W31130173131 @default.
- W3113017313 hasLocation W31130173132 @default.
- W3113017313 hasLocation W31130173133 @default.
- W3113017313 hasOpenAccess W3113017313 @default.
- W3113017313 hasPrimaryLocation W31130173131 @default.
- W3113017313 hasRelatedWork W2040397200 @default.
- W3113017313 hasRelatedWork W2137463787 @default.
- W3113017313 hasRelatedWork W2365213443 @default.
- W3113017313 hasRelatedWork W2748454020 @default.
- W3113017313 hasRelatedWork W2767651786 @default.
- W3113017313 hasRelatedWork W2773500201 @default.
- W3113017313 hasRelatedWork W2912288872 @default.
- W3113017313 hasRelatedWork W3119610945 @default.
- W3113017313 hasRelatedWork W4384103574 @default.
- W3113017313 hasRelatedWork W564581980 @default.
- W3113017313 hasVolume "14" @default.
- W3113017313 isParatext "false" @default.