Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912173254> ?p ?o ?g. }
- W2912173254 abstract "Deep neural networks achieve stellar generalisation on a variety of problems, despite often being large enough to easily fit all their training data. Here we study the generalisation dynamics of two-layer neural networks in a teacher-student setup, where one network, the student, is trained using stochastic gradient descent (SGD) on data generated by another network, called the teacher. We show how for this problem, the dynamics of SGD are captured by a set of differential equations. In particular, we demonstrate analytically that the generalisation error of the student increases linearly with the network size, with other relevant parameters held constant. Our results indicate that achieving good generalisation in neural networks depends on the interplay of at least the algorithm, its learning rate, the model architecture, and the data set." @default.
- W2912173254 created "2019-02-21" @default.
- W2912173254 creator A5011428379 @default.
- W2912173254 creator A5042904819 @default.
- W2912173254 creator A5057039281 @default.
- W2912173254 creator A5068236230 @default.
- W2912173254 creator A5089268172 @default.
- W2912173254 date "2019-01-25" @default.
- W2912173254 modified "2023-10-17" @default.
- W2912173254 title "Generalisation dynamics of online learning in over-parameterised neural networks" @default.
- W2912173254 cites W1488975066 @default.
- W2912173254 cites W1944672 @default.
- W2912173254 cites W1964862779 @default.
- W2912173254 cites W1995842804 @default.
- W2912173254 cites W2006997461 @default.
- W2912173254 cites W2037985840 @default.
- W2912173254 cites W2042318263 @default.
- W2912173254 cites W2050583479 @default.
- W2912173254 cites W2066424095 @default.
- W2912173254 cites W2081748565 @default.
- W2912173254 cites W2090614046 @default.
- W2912173254 cites W2103496339 @default.
- W2912173254 cites W2150872430 @default.
- W2912173254 cites W2257979135 @default.
- W2912173254 cites W2579923771 @default.
- W2912173254 cites W2763894180 @default.
- W2912173254 cites W2772785876 @default.
- W2912173254 cites W2777138330 @default.
- W2912173254 cites W2804589149 @default.
- W2912173254 cites W2804639765 @default.
- W2912173254 cites W2886067286 @default.
- W2912173254 cites W2903327037 @default.
- W2912173254 cites W2919115771 @default.
- W2912173254 cites W2962835968 @default.
- W2912173254 cites W2962857907 @default.
- W2912173254 cites W2962973336 @default.
- W2912173254 cites W2963095610 @default.
- W2912173254 cites W2963096987 @default.
- W2912173254 cites W2963100491 @default.
- W2912173254 cites W2963201159 @default.
- W2912173254 cites W2963236897 @default.
- W2912173254 cites W2963417959 @default.
- W2912173254 cites W2963504252 @default.
- W2912173254 cites W2963509076 @default.
- W2912173254 cites W2963695615 @default.
- W2912173254 cites W2964156139 @default.
- W2912173254 cites W2970330753 @default.
- W2912173254 cites W3093329015 @default.
- W2912173254 cites W3104787172 @default.
- W2912173254 cites W3137695714 @default.
- W2912173254 cites W3146803896 @default.
- W2912173254 cites W3141350557 @default.
- W2912173254 hasPublicationYear "2019" @default.
- W2912173254 type Work @default.
- W2912173254 sameAs 2912173254 @default.
- W2912173254 citedByCount "7" @default.
- W2912173254 countsByYear W29121732542018 @default.
- W2912173254 countsByYear W29121732542019 @default.
- W2912173254 countsByYear W29121732542020 @default.
- W2912173254 countsByYear W29121732542021 @default.
- W2912173254 crossrefType "posted-content" @default.
- W2912173254 hasAuthorship W2912173254A5011428379 @default.
- W2912173254 hasAuthorship W2912173254A5042904819 @default.
- W2912173254 hasAuthorship W2912173254A5057039281 @default.
- W2912173254 hasAuthorship W2912173254A5068236230 @default.
- W2912173254 hasAuthorship W2912173254A5089268172 @default.
- W2912173254 hasConcept C119857082 @default.
- W2912173254 hasConcept C121332964 @default.
- W2912173254 hasConcept C136197465 @default.
- W2912173254 hasConcept C145912823 @default.
- W2912173254 hasConcept C153258448 @default.
- W2912173254 hasConcept C154945302 @default.
- W2912173254 hasConcept C175202392 @default.
- W2912173254 hasConcept C177264268 @default.
- W2912173254 hasConcept C199360897 @default.
- W2912173254 hasConcept C206688291 @default.
- W2912173254 hasConcept C24890656 @default.
- W2912173254 hasConcept C2777027219 @default.
- W2912173254 hasConcept C41008148 @default.
- W2912173254 hasConcept C50644808 @default.
- W2912173254 hasConcept C58489278 @default.
- W2912173254 hasConcept C86582703 @default.
- W2912173254 hasConceptScore W2912173254C119857082 @default.
- W2912173254 hasConceptScore W2912173254C121332964 @default.
- W2912173254 hasConceptScore W2912173254C136197465 @default.
- W2912173254 hasConceptScore W2912173254C145912823 @default.
- W2912173254 hasConceptScore W2912173254C153258448 @default.
- W2912173254 hasConceptScore W2912173254C154945302 @default.
- W2912173254 hasConceptScore W2912173254C175202392 @default.
- W2912173254 hasConceptScore W2912173254C177264268 @default.
- W2912173254 hasConceptScore W2912173254C199360897 @default.
- W2912173254 hasConceptScore W2912173254C206688291 @default.
- W2912173254 hasConceptScore W2912173254C24890656 @default.
- W2912173254 hasConceptScore W2912173254C2777027219 @default.
- W2912173254 hasConceptScore W2912173254C41008148 @default.
- W2912173254 hasConceptScore W2912173254C50644808 @default.
- W2912173254 hasConceptScore W2912173254C58489278 @default.
- W2912173254 hasConceptScore W2912173254C86582703 @default.
- W2912173254 hasOpenAccess W2912173254 @default.
- W2912173254 hasRelatedWork W2153814050 @default.