FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4317438807> ?p ?o ?g. }

• W4317438807 endingPage "e1010808" @default.
• W4317438807 startingPage "e1010808" @default.
• W4317438807 abstract "Humans can learn several tasks in succession with minimal mutual interference but perform more poorly when trained on multiple tasks at once. The opposite is true for standard deep neural networks. Here, we propose novel computational constraints for artificial neural networks, inspired by earlier work on gating in the primate prefrontal cortex, that capture the cost of interleaved training and allow the network to learn two tasks in sequence without forgetting. We augment standard stochastic gradient descent with two algorithmic motifs, so-called “sluggish” task units and a Hebbian training step that strengthens connections between task units and hidden units that encode task-relevant information. We found that the “sluggish” units introduce a switch-cost during training, which biases representations under interleaved training towards a joint representation that ignores the contextual cue, while the Hebbian step promotes the formation of a gating scheme from task units to the hidden layer that produces orthogonal representations which are perfectly guarded against interference. Validating the model on previously published human behavioural data revealed that it matches performance of participants who had been trained on blocked or interleaved curricula, and that these performance differences were driven by misestimation of the true category boundary." @default.
• W4317438807 created "2023-01-19" @default.
• W4317438807 creator A5011428379 @default.
• W4317438807 creator A5031878516 @default.
• W4317438807 creator A5054786733 @default.
• W4317438807 creator A5084904199 @default.
• W4317438807 date "2023-01-19" @default.
• W4317438807 modified "2023-10-18" @default.
• W4317438807 title "Modelling continual learning in humans with Hebbian context gating and exponentially decaying task signals" @default.
• W4317438807 cites W1715013381 @default.
• W4317438807 cites W1964785871 @default.
• W4317438807 cites W1973967428 @default.
• W4317438807 cites W1996538430 @default.
• W4317438807 cites W2037944537 @default.
• W4317438807 cites W2040036684 @default.
• W4317438807 cites W2041159118 @default.
• W4317438807 cites W2047057213 @default.
• W4317438807 cites W2047125104 @default.
• W4317438807 cites W2050583479 @default.
• W4317438807 cites W2058616551 @default.
• W4317438807 cites W2060277733 @default.
• W4317438807 cites W2060618570 @default.
• W4317438807 cites W2061562057 @default.
• W4317438807 cites W2061832492 @default.
• W4317438807 cites W2083133776 @default.
• W4317438807 cites W2105466163 @default.
• W4317438807 cites W2107564936 @default.
• W4317438807 cites W2111534506 @default.
• W4317438807 cites W2116640775 @default.
• W4317438807 cites W2121102898 @default.
• W4317438807 cites W2122917065 @default.
• W4317438807 cites W2125495031 @default.
• W4317438807 cites W2145339207 @default.
• W4317438807 cites W2151137320 @default.
• W4317438807 cites W2152382177 @default.
• W4317438807 cites W2432567885 @default.
• W4317438807 cites W2529605558 @default.
• W4317438807 cites W2560647685 @default.
• W4317438807 cites W2737492962 @default.
• W4317438807 cites W2753709519 @default.
• W4317438807 cites W2787295326 @default.
• W4317438807 cites W2788388592 @default.
• W4317438807 cites W2896084542 @default.
• W4317438807 cites W2899696887 @default.
• W4317438807 cites W2900440811 @default.
• W4317438807 cites W2908124316 @default.
• W4317438807 cites W2967821093 @default.
• W4317438807 cites W2978368159 @default.
• W4317438807 cites W3004895274 @default.
• W4317438807 cites W3016035772 @default.
• W4317438807 cites W3048941751 @default.
• W4317438807 cites W3081847250 @default.
• W4317438807 cites W3096511324 @default.
• W4317438807 cites W3097816393 @default.
• W4317438807 cites W3100035092 @default.
• W4317438807 cites W3119948327 @default.
• W4317438807 cites W3140601242 @default.
• W4317438807 cites W3143658323 @default.
• W4317438807 cites W3161865842 @default.
• W4317438807 cites W3211851621 @default.
• W4317438807 cites W4200450018 @default.
• W4317438807 cites W4211115742 @default.
• W4317438807 cites W4225131785 @default.
• W4317438807 cites W4225585129 @default.
• W4317438807 cites W4238879677 @default.
• W4317438807 cites W4312018460 @default.
• W4317438807 cites W4312156980 @default.
• W4317438807 doi "https://doi.org/10.1371/journal.pcbi.1010808" @default.
• W4317438807 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/36656823" @default.
• W4317438807 hasPublicationYear "2023" @default.
• W4317438807 type Work @default.
• W4317438807 citedByCount "7" @default.
• W4317438807 countsByYear W43174388072022 @default.
• W4317438807 countsByYear W43174388072023 @default.
• W4317438807 crossrefType "journal-article" @default.
• W4317438807 hasAuthorship W4317438807A5011428379 @default.
• W4317438807 hasAuthorship W4317438807A5031878516 @default.
• W4317438807 hasAuthorship W4317438807A5054786733 @default.
• W4317438807 hasAuthorship W4317438807A5084904199 @default.
• W4317438807 hasBestOaLocation W43174388071 @default.
• W4317438807 hasConcept C111437709 @default.
• W4317438807 hasConcept C111919701 @default.
• W4317438807 hasConcept C119857082 @default.
• W4317438807 hasConcept C127705205 @default.
• W4317438807 hasConcept C151730666 @default.
• W4317438807 hasConcept C153180895 @default.
• W4317438807 hasConcept C154945302 @default.
• W4317438807 hasConcept C15744967 @default.
• W4317438807 hasConcept C162324750 @default.
• W4317438807 hasConcept C180747234 @default.
• W4317438807 hasConcept C187736073 @default.
• W4317438807 hasConcept C194544171 @default.
• W4317438807 hasConcept C2779343474 @default.
• W4317438807 hasConcept C2780451532 @default.
• W4317438807 hasConcept C41008148 @default.
• W4317438807 hasConcept C42407357 @default.
• W4317438807 hasConcept C46421273 @default.
• W4317438807 hasConcept C50644808 @default.

• next