FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3179888904> ?p ?o ?g. }

• W3179888904 endingPage "21" @default.
• W3179888904 startingPage "1" @default.
• W3179888904 abstract "Abstract Deep neural networks (DNNs) trained on object recognition provide the best current models of high-level visual cortex. What remains unclear is how strongly experimental choices, such as network architecture, training, and fitting to brain data, contribute to the observed similarities. Here, we compare a diverse set of nine DNN architectures on their ability to explain the representational geometry of 62 object images in human inferior temporal cortex (hIT), as measured with fMRI. We compare untrained networks to their task-trained counterparts and assess the effect of cross-validated fitting to hIT, by taking a weighted combination of the principal components of features within each layer and, subsequently, a weighted combination of layers. For each combination of training and fitting, we test all models for their correlation with the hIT representational dissimilarity matrix, using independent images and subjects. Trained models outperform untrained models (accounting for 57% more of the explainable variance), suggesting that structured visual features are important for explaining hIT. Model fitting further improves the alignment of DNN and hIT representations (by 124%), suggesting that the relative prevalence of different features in hIT does not readily emerge from the Imagenet object-recognition task used to train the networks. The same models can also explain the disparate representations in primary visual cortex (V1), where stronger weights are given to earlier layers. In each region, all architectures achieved equivalently high performance once trained and fitted. The models' shared properties—deep feedforward hierarchies of spatially restricted nonlinear filters—seem more important than their differences, when modeling human visual representations." @default.
• W3179888904 created "2021-07-19" @default.
• W3179888904 creator A5021717958 @default.
• W3179888904 creator A5024544588 @default.
• W3179888904 creator A5042387485 @default.
• W3179888904 creator A5046328619 @default.
• W3179888904 creator A5084467223 @default.
• W3179888904 date "2021-08-19" @default.
• W3179888904 modified "2023-09-29" @default.
• W3179888904 title "Diverse Deep Neural Networks All Predict Human Inferior Temporal Cortex Well, After Training and Fitting" @default.
• W3179888904 cites W144689751 @default.
• W3179888904 cites W1686810756 @default.
• W3179888904 cites W1715013381 @default.
• W3179888904 cites W1970167670 @default.
• W3179888904 cites W197865394 @default.
• W3179888904 cites W2000908757 @default.
• W3179888904 cites W2032033817 @default.
• W3179888904 cites W2040036684 @default.
• W3179888904 cites W2057307785 @default.
• W3179888904 cites W2058616551 @default.
• W3179888904 cites W2116360511 @default.
• W3179888904 cites W2117539524 @default.
• W3179888904 cites W2123544472 @default.
• W3179888904 cites W2149194912 @default.
• W3179888904 cites W2160654481 @default.
• W3179888904 cites W2163605009 @default.
• W3179888904 cites W2166206801 @default.
• W3179888904 cites W2176287621 @default.
• W3179888904 cites W2220779336 @default.
• W3179888904 cites W2274405424 @default.
• W3179888904 cites W2279098554 @default.
• W3179888904 cites W2337199865 @default.
• W3179888904 cites W2340573780 @default.
• W3179888904 cites W2412480261 @default.
• W3179888904 cites W2494117141 @default.
• W3179888904 cites W2537084945 @default.
• W3179888904 cites W2568719307 @default.
• W3179888904 cites W2610765860 @default.
• W3179888904 cites W2611715315 @default.
• W3179888904 cites W2662969263 @default.
• W3179888904 cites W2762409611 @default.
• W3179888904 cites W2763767712 @default.
• W3179888904 cites W2797270586 @default.
• W3179888904 cites W2892147425 @default.
• W3179888904 cites W2898929289 @default.
• W3179888904 cites W2943083682 @default.
• W3179888904 cites W2951065015 @default.
• W3179888904 cites W2951506741 @default.
• W3179888904 cites W2952116081 @default.
• W3179888904 cites W2952828126 @default.
• W3179888904 cites W2963341661 @default.
• W3179888904 cites W2964350391 @default.
• W3179888904 cites W2978368159 @default.
• W3179888904 cites W2978664050 @default.
• W3179888904 cites W2979357328 @default.
• W3179888904 cites W3004895274 @default.
• W3179888904 cites W3006472786 @default.
• W3179888904 cites W3013386261 @default.
• W3179888904 cites W3013414436 @default.
• W3179888904 cites W3016503290 @default.
• W3179888904 cites W3016970897 @default.
• W3179888904 cites W3038401053 @default.
• W3179888904 cites W3040750741 @default.
• W3179888904 cites W3041318187 @default.
• W3179888904 cites W3092472394 @default.
• W3179888904 cites W3100481436 @default.
• W3179888904 cites W3101378249 @default.
• W3179888904 cites W3108490213 @default.
• W3179888904 cites W3127176114 @default.
• W3179888904 cites W3129520418 @default.
• W3179888904 cites W3129717984 @default.
• W3179888904 cites W3137042839 @default.
• W3179888904 cites W3146620971 @default.
• W3179888904 cites W3157521173 @default.
• W3179888904 cites W3194437088 @default.
• W3179888904 doi "https://doi.org/10.1162/jocn_a_01755" @default.
• W3179888904 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/34272948" @default.
• W3179888904 hasPublicationYear "2021" @default.
• W3179888904 type Work @default.
• W3179888904 sameAs 3179888904 @default.
• W3179888904 citedByCount "28" @default.
• W3179888904 countsByYear W31798889042021 @default.
• W3179888904 countsByYear W31798889042022 @default.
• W3179888904 countsByYear W31798889042023 @default.
• W3179888904 crossrefType "journal-article" @default.
• W3179888904 hasAuthorship W3179888904A5021717958 @default.
• W3179888904 hasAuthorship W3179888904A5024544588 @default.
• W3179888904 hasAuthorship W3179888904A5042387485 @default.
• W3179888904 hasAuthorship W3179888904A5046328619 @default.
• W3179888904 hasAuthorship W3179888904A5084467223 @default.
• W3179888904 hasBestOaLocation W31798889042 @default.
• W3179888904 hasConcept C119857082 @default.
• W3179888904 hasConcept C121955636 @default.
• W3179888904 hasConcept C144133560 @default.
• W3179888904 hasConcept C153180895 @default.
• W3179888904 hasConcept C154945302 @default.
• W3179888904 hasConcept C15744967 @default.
• W3179888904 hasConcept C162324750 @default.

• next