Matches in SemOpenAlex for { <https://semopenalex.org/work/W3146058829> ?p ?o ?g. }
- W3146058829 endingPage "293" @default.
- W3146058829 startingPage "278" @default.
- W3146058829 abstract "We introduce the Generalized Discrimination Value (GDV) that measures, in a non-invasive manner, how well different data classes separate in each given layer of an artificial neural network. It turns out that, at the end of the training period, the GDV in each given layer L attains a highly reproducible value, irrespective of the initialization of the network’s connection weights. In the case of multi-layer perceptrons trained with error backpropagation, we find that classification of highly complex data sets requires a temporal reduction of class separability, marked by a characteristic ‘energy barrier’ in the initial part of the GDV(L) curve. Even more surprisingly, for a given data set, the GDV(L) is running through a fixed ‘master curve’, independently from the total number of network layers. Finally, due to its invariance with respect to dimensionality, the GDV may serve as a useful tool to compare the internal representational dynamics of artificial neural networks with different architectures for neural architecture search or network compression; or even with brain activity in order to decide between different candidate models of brain function." @default.
- W3146058829 created "2021-04-13" @default.
- W3146058829 creator A5009671657 @default.
- W3146058829 creator A5026690582 @default.
- W3146058829 creator A5032499810 @default.
- W3146058829 creator A5049878919 @default.
- W3146058829 creator A5062539159 @default.
- W3146058829 date "2021-07-01" @default.
- W3146058829 modified "2023-10-16" @default.
- W3146058829 title "Quantifying the separability of data classes in neural networks" @default.
- W3146058829 cites W143174683 @default.
- W3146058829 cites W1972611943 @default.
- W3146058829 cites W1973192023 @default.
- W3146058829 cites W1993436046 @default.
- W3146058829 cites W2003386389 @default.
- W3146058829 cites W2057307785 @default.
- W3146058829 cites W2064675550 @default.
- W3146058829 cites W2078339916 @default.
- W3146058829 cites W2095293504 @default.
- W3146058829 cites W2112796928 @default.
- W3146058829 cites W2145339207 @default.
- W3146058829 cites W2146292423 @default.
- W3146058829 cites W2151936673 @default.
- W3146058829 cites W2166049352 @default.
- W3146058829 cites W2176287621 @default.
- W3146058829 cites W2183341477 @default.
- W3146058829 cites W2194775991 @default.
- W3146058829 cites W2252909801 @default.
- W3146058829 cites W2257979135 @default.
- W3146058829 cites W2531409750 @default.
- W3146058829 cites W2542768043 @default.
- W3146058829 cites W2576404523 @default.
- W3146058829 cites W2589063200 @default.
- W3146058829 cites W2622826443 @default.
- W3146058829 cites W2726198599 @default.
- W3146058829 cites W2738724892 @default.
- W3146058829 cites W2755036008 @default.
- W3146058829 cites W2762466482 @default.
- W3146058829 cites W2766362701 @default.
- W3146058829 cites W2766447205 @default.
- W3146058829 cites W2778838066 @default.
- W3146058829 cites W2779578326 @default.
- W3146058829 cites W2787225861 @default.
- W3146058829 cites W2792641098 @default.
- W3146058829 cites W2795176995 @default.
- W3146058829 cites W2800142021 @default.
- W3146058829 cites W2883583109 @default.
- W3146058829 cites W2884430236 @default.
- W3146058829 cites W2886490392 @default.
- W3146058829 cites W2887972576 @default.
- W3146058829 cites W2901623423 @default.
- W3146058829 cites W2908201961 @default.
- W3146058829 cites W2919115771 @default.
- W3146058829 cites W2950987652 @default.
- W3146058829 cites W2951065015 @default.
- W3146058829 cites W2951877485 @default.
- W3146058829 cites W2962956675 @default.
- W3146058829 cites W2963556566 @default.
- W3146058829 cites W2964081807 @default.
- W3146058829 cites W2971337080 @default.
- W3146058829 cites W2993475131 @default.
- W3146058829 cites W2993894543 @default.
- W3146058829 cites W3014541266 @default.
- W3146058829 cites W3040625097 @default.
- W3146058829 cites W3099305384 @default.
- W3146058829 cites W3105432754 @default.
- W3146058829 cites W3125908253 @default.
- W3146058829 cites W4229494842 @default.
- W3146058829 cites W2936855705 @default.
- W3146058829 doi "https://doi.org/10.1016/j.neunet.2021.03.035" @default.
- W3146058829 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/33862387" @default.
- W3146058829 hasPublicationYear "2021" @default.
- W3146058829 type Work @default.
- W3146058829 sameAs 3146058829 @default.
- W3146058829 citedByCount "18" @default.
- W3146058829 countsByYear W31460588292021 @default.
- W3146058829 countsByYear W31460588292022 @default.
- W3146058829 countsByYear W31460588292023 @default.
- W3146058829 crossrefType "journal-article" @default.
- W3146058829 hasAuthorship W3146058829A5009671657 @default.
- W3146058829 hasAuthorship W3146058829A5026690582 @default.
- W3146058829 hasAuthorship W3146058829A5032499810 @default.
- W3146058829 hasAuthorship W3146058829A5049878919 @default.
- W3146058829 hasAuthorship W3146058829A5062539159 @default.
- W3146058829 hasBestOaLocation W31460588291 @default.
- W3146058829 hasConcept C11413529 @default.
- W3146058829 hasConcept C114466953 @default.
- W3146058829 hasConcept C153180895 @default.
- W3146058829 hasConcept C154945302 @default.
- W3146058829 hasConcept C155032097 @default.
- W3146058829 hasConcept C177264268 @default.
- W3146058829 hasConcept C178790620 @default.
- W3146058829 hasConcept C179717631 @default.
- W3146058829 hasConcept C185592680 @default.
- W3146058829 hasConcept C199360897 @default.
- W3146058829 hasConcept C202286095 @default.
- W3146058829 hasConcept C2779227376 @default.
- W3146058829 hasConcept C41008148 @default.