Matches in SemOpenAlex for { <https://semopenalex.org/work/W3102618132> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W3102618132 endingPage "7307" @default.
- W3102618132 startingPage "7296" @default.
- W3102618132 abstract "The state-of-the art machine learning approach to training deep neural networks, backpropagation, is implausible for real neural networks: neurons need to know their outgoing weights; training alternates between a bottom-up forward pass (computation) and a top-down backward pass (learning); and the algorithm often needs precise labels of many data points. Biologically plausible approximations to backpropagation, such as feedback alignment, solve the weight transport problem, but not the other two. Thus, fully biologically plausible learning rules have so far remained elusive. Here we present a family of learning rules that does not suffer from any of these problems. It is motivated by the information bottleneck principle (extended with kernel methods), in which networks learn to compress the input as much as possible without sacrificing prediction of the output. The resulting rules have a 3-factor Hebbian structure: they require pre- and post-synaptic firing rates and an error signal - the third factor - consisting of a global teaching signal and a layer-specific term, both available without a top-down pass. They do not require precise labels; instead, they rely on the similarity between pairs of desired outputs. Moreover, to obtain good performance on hard problems and retain biological plausibility, our rules need divisive normalization - a known feature of biological networks. Finally, simulations show that our rules perform nearly as well as backpropagation on image classification tasks." @default.
- W3102618132 created "2020-11-23" @default.
- W3102618132 creator A5025338734 @default.
- W3102618132 creator A5027156456 @default.
- W3102618132 date "2020-01-01" @default.
- W3102618132 modified "2023-09-24" @default.
- W3102618132 title "Kernelized information bottleneck leads to biologically plausible 3-factor Hebbian learning in deep networks" @default.
- W3102618132 hasPublicationYear "2020" @default.
- W3102618132 type Work @default.
- W3102618132 sameAs 3102618132 @default.
- W3102618132 citedByCount "3" @default.
- W3102618132 countsByYear W31026181322021 @default.
- W3102618132 crossrefType "proceedings-article" @default.
- W3102618132 hasAuthorship W3102618132A5025338734 @default.
- W3102618132 hasAuthorship W3102618132A5027156456 @default.
- W3102618132 hasConcept C108583219 @default.
- W3102618132 hasConcept C111437709 @default.
- W3102618132 hasConcept C114614502 @default.
- W3102618132 hasConcept C117765406 @default.
- W3102618132 hasConcept C119857082 @default.
- W3102618132 hasConcept C136886441 @default.
- W3102618132 hasConcept C144024400 @default.
- W3102618132 hasConcept C149635348 @default.
- W3102618132 hasConcept C153180895 @default.
- W3102618132 hasConcept C154945302 @default.
- W3102618132 hasConcept C155032097 @default.
- W3102618132 hasConcept C17061570 @default.
- W3102618132 hasConcept C19165224 @default.
- W3102618132 hasConcept C2779127903 @default.
- W3102618132 hasConcept C2780513914 @default.
- W3102618132 hasConcept C33923547 @default.
- W3102618132 hasConcept C41008148 @default.
- W3102618132 hasConcept C50644808 @default.
- W3102618132 hasConcept C60008888 @default.
- W3102618132 hasConcept C73555534 @default.
- W3102618132 hasConcept C74193536 @default.
- W3102618132 hasConcept C97108695 @default.
- W3102618132 hasConceptScore W3102618132C108583219 @default.
- W3102618132 hasConceptScore W3102618132C111437709 @default.
- W3102618132 hasConceptScore W3102618132C114614502 @default.
- W3102618132 hasConceptScore W3102618132C117765406 @default.
- W3102618132 hasConceptScore W3102618132C119857082 @default.
- W3102618132 hasConceptScore W3102618132C136886441 @default.
- W3102618132 hasConceptScore W3102618132C144024400 @default.
- W3102618132 hasConceptScore W3102618132C149635348 @default.
- W3102618132 hasConceptScore W3102618132C153180895 @default.
- W3102618132 hasConceptScore W3102618132C154945302 @default.
- W3102618132 hasConceptScore W3102618132C155032097 @default.
- W3102618132 hasConceptScore W3102618132C17061570 @default.
- W3102618132 hasConceptScore W3102618132C19165224 @default.
- W3102618132 hasConceptScore W3102618132C2779127903 @default.
- W3102618132 hasConceptScore W3102618132C2780513914 @default.
- W3102618132 hasConceptScore W3102618132C33923547 @default.
- W3102618132 hasConceptScore W3102618132C41008148 @default.
- W3102618132 hasConceptScore W3102618132C50644808 @default.
- W3102618132 hasConceptScore W3102618132C60008888 @default.
- W3102618132 hasConceptScore W3102618132C73555534 @default.
- W3102618132 hasConceptScore W3102618132C74193536 @default.
- W3102618132 hasConceptScore W3102618132C97108695 @default.
- W3102618132 hasLocation W31026181321 @default.
- W3102618132 hasOpenAccess W3102618132 @default.
- W3102618132 hasPrimaryLocation W31026181321 @default.
- W3102618132 hasRelatedWork W2062573803 @default.
- W3102618132 hasRelatedWork W2117933835 @default.
- W3102618132 hasRelatedWork W2147368400 @default.
- W3102618132 hasRelatedWork W2156938623 @default.
- W3102618132 hasRelatedWork W2182969463 @default.
- W3102618132 hasRelatedWork W2355715145 @default.
- W3102618132 hasRelatedWork W2786262858 @default.
- W3102618132 hasRelatedWork W2786465559 @default.
- W3102618132 hasRelatedWork W2898807633 @default.
- W3102618132 hasRelatedWork W2952569483 @default.
- W3102618132 hasRelatedWork W2953042818 @default.
- W3102618132 hasRelatedWork W3006378306 @default.
- W3102618132 hasRelatedWork W3033626776 @default.
- W3102618132 hasRelatedWork W3035170382 @default.
- W3102618132 hasRelatedWork W3126611227 @default.
- W3102618132 hasRelatedWork W3132798516 @default.
- W3102618132 hasRelatedWork W3174753037 @default.
- W3102618132 hasRelatedWork W3212945014 @default.
- W3102618132 hasRelatedWork W772458362 @default.
- W3102618132 hasRelatedWork W3093873179 @default.
- W3102618132 hasVolume "33" @default.
- W3102618132 isParatext "false" @default.
- W3102618132 isRetracted "false" @default.
- W3102618132 magId "3102618132" @default.
- W3102618132 workType "article" @default.