FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2070922095> ?p ?o ?g. }

• W2070922095 endingPage "15" @default.
• W2070922095 startingPage "6" @default.
• W2070922095 abstract "In this paper, we elaborate upon the claim that clustering in the recurrent layer of recurrent neural networks (RNNs) reflects meaningful information processing states even prior to training. By concentrating on activation clusters in RNNs, while not throwing away the continuous state space network dynamics, we extract predictive models that we call neural prediction machines (NPMs). When RNNs with sigmoid activation functions are initialized with small weights (a common technique in the RNN community), the clusters of recurrent activations emerging prior to training are indeed meaningful and correspond to Markov prediction contexts. In this case, the extracted NPMs correspond to a class of Markov models, called variable memory length Markov models (VLMMs). In order to appreciate how much information has really been induced during the training, the RNN performance should always be compared with that of VLMMs and NPMs extracted before training as the null base models. Our arguments are supported by experiments on a chaotic symbolic sequence and a context-free language with a deep recursive structure." @default.
• W2070922095 created "2016-06-24" @default.
• W2070922095 creator A5002021863 @default.
• W2070922095 creator A5034079531 @default.
• W2070922095 creator A5051182859 @default.
• W2070922095 date "2004-01-01" @default.
• W2070922095 modified "2023-10-18" @default.
• W2070922095 title "Markovian Architectural Bias of Recurrent Neural Networks" @default.
• W2070922095 cites W1592855356 @default.
• W2070922095 cites W1839273017 @default.
• W2070922095 cites W1965511524 @default.
• W2070922095 cites W1972085588 @default.
• W2070922095 cites W1982370770 @default.
• W2070922095 cites W1997810540 @default.
• W2070922095 cites W2000309080 @default.
• W2070922095 cites W2010315176 @default.
• W2070922095 cites W2011485152 @default.
• W2070922095 cites W2016589492 @default.
• W2070922095 cites W2053987251 @default.
• W2070922095 cites W2063370840 @default.
• W2070922095 cites W2064675550 @default.
• W2070922095 cites W2076332735 @default.
• W2070922095 cites W2077439714 @default.
• W2070922095 cites W2082345008 @default.
• W2070922095 cites W2082967074 @default.
• W2070922095 cites W20854365 @default.
• W2070922095 cites W2095963120 @default.
• W2070922095 cites W2098038521 @default.
• W2070922095 cites W2102834748 @default.
• W2070922095 cites W2107878631 @default.
• W2070922095 cites W2117042439 @default.
• W2070922095 cites W2120339295 @default.
• W2070922095 cites W2121553911 @default.
• W2070922095 cites W2133656308 @default.
• W2070922095 cites W2133688510 @default.
• W2070922095 cites W2134795071 @default.
• W2070922095 cites W2136451213 @default.
• W2070922095 cites W2144001972 @default.
• W2070922095 cites W2144166812 @default.
• W2070922095 cites W2150039113 @default.
• W2070922095 cites W2150355110 @default.
• W2070922095 cites W2156960699 @default.
• W2070922095 cites W2157127400 @default.
• W2070922095 cites W2169791674 @default.
• W2070922095 cites W4212958445 @default.
• W2070922095 cites W4247844322 @default.
• W2070922095 cites W4254816979 @default.
• W2070922095 doi "https://doi.org/10.1109/tnn.2003.820839" @default.
• W2070922095 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/15387243" @default.
• W2070922095 hasPublicationYear "2004" @default.
• W2070922095 type Work @default.
• W2070922095 sameAs 2070922095 @default.
• W2070922095 citedByCount "158" @default.
• W2070922095 countsByYear W20709220952012 @default.
• W2070922095 countsByYear W20709220952013 @default.
• W2070922095 countsByYear W20709220952014 @default.
• W2070922095 countsByYear W20709220952015 @default.
• W2070922095 countsByYear W20709220952016 @default.
• W2070922095 countsByYear W20709220952017 @default.
• W2070922095 countsByYear W20709220952018 @default.
• W2070922095 countsByYear W20709220952019 @default.
• W2070922095 countsByYear W20709220952020 @default.
• W2070922095 countsByYear W20709220952021 @default.
• W2070922095 countsByYear W20709220952022 @default.
• W2070922095 countsByYear W20709220952023 @default.
• W2070922095 crossrefType "journal-article" @default.
• W2070922095 hasAuthorship W2070922095A5002021863 @default.
• W2070922095 hasAuthorship W2070922095A5034079531 @default.
• W2070922095 hasAuthorship W2070922095A5051182859 @default.
• W2070922095 hasConcept C105795698 @default.
• W2070922095 hasConcept C119857082 @default.
• W2070922095 hasConcept C147168706 @default.
• W2070922095 hasConcept C151730666 @default.
• W2070922095 hasConcept C154945302 @default.
• W2070922095 hasConcept C159886148 @default.
• W2070922095 hasConcept C163836022 @default.
• W2070922095 hasConcept C23224414 @default.
• W2070922095 hasConcept C2779343474 @default.
• W2070922095 hasConcept C33923547 @default.
• W2070922095 hasConcept C41008148 @default.
• W2070922095 hasConcept C50644808 @default.
• W2070922095 hasConcept C72434380 @default.
• W2070922095 hasConcept C73555534 @default.
• W2070922095 hasConcept C81388566 @default.
• W2070922095 hasConcept C86803240 @default.
• W2070922095 hasConcept C98763669 @default.
• W2070922095 hasConceptScore W2070922095C105795698 @default.
• W2070922095 hasConceptScore W2070922095C119857082 @default.
• W2070922095 hasConceptScore W2070922095C147168706 @default.
• W2070922095 hasConceptScore W2070922095C151730666 @default.
• W2070922095 hasConceptScore W2070922095C154945302 @default.
• W2070922095 hasConceptScore W2070922095C159886148 @default.
• W2070922095 hasConceptScore W2070922095C163836022 @default.
• W2070922095 hasConceptScore W2070922095C23224414 @default.
• W2070922095 hasConceptScore W2070922095C2779343474 @default.
• W2070922095 hasConceptScore W2070922095C33923547 @default.
• W2070922095 hasConceptScore W2070922095C41008148 @default.
• W2070922095 hasConceptScore W2070922095C50644808 @default.

• next