Matches in SemOpenAlex for { <https://semopenalex.org/work/W2296394872> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W2296394872 abstract "Catastrophic forgetting is a problem encountered with neural networks as well as other learning systems whereby past representations are lost as new representations are learned. It has been shown that catastrophic forgetting can be mitigated in neural networks by using a neuron selection technique, dubbed cluster-select, which performs online clustering over the network inputs to partition the network such that only a subset of neurons are used for a given input vector. Cluster-select can benefit by using Mahalanobis distance which relies on an inverse covariance estimate. Unfortunately, covariance estimation is problematic when lacking a very large number of samples relative to the number of input dimensions. One way to tackle this problem is through the use of a shrinkage estimator that offers a compromise between the sample covariance matrix and a well-conditioned matrix with the aim of minimizing the mean-squared error (MSE). In online environments, such as those in which catastrophic forgetting can occur, data arrives sequentially, requiring the covariance matrix to be estimated sequentially. Therefore, in this work we derive sequential update rules for the shrinkage estimator and approximate it's related inverse. The online covariance estimator is applied to the cluster-select technique with results that demonstrate further improvements in terms of effectively mitigating catastrophic forgetting." @default.
- W2296394872 created "2016-06-24" @default.
- W2296394872 creator A5025477216 @default.
- W2296394872 creator A5035780973 @default.
- W2296394872 creator A5081192463 @default.
- W2296394872 date "2015-12-01" @default.
- W2296394872 modified "2023-09-27" @default.
- W2296394872 title "Sequential Covariance-Matrix Estimation with Application to Mitigating Catastrophic Forgetting" @default.
- W2296394872 cites W1584444527 @default.
- W2296394872 cites W1598266570 @default.
- W2296394872 cites W1925816294 @default.
- W2296394872 cites W1969885712 @default.
- W2296394872 cites W1976370317 @default.
- W2296394872 cites W1980326979 @default.
- W2296394872 cites W1985291814 @default.
- W2296394872 cites W1990512452 @default.
- W2296394872 cites W1991289661 @default.
- W2296394872 cites W1996340400 @default.
- W2296394872 cites W2013122247 @default.
- W2296394872 cites W2035729201 @default.
- W2296394872 cites W2045490871 @default.
- W2296394872 cites W2055701312 @default.
- W2296394872 cites W2062125287 @default.
- W2296394872 cites W2099049980 @default.
- W2296394872 cites W2100495367 @default.
- W2296394872 cites W2113207845 @default.
- W2296394872 cites W2113839990 @default.
- W2296394872 cites W2145339207 @default.
- W2296394872 cites W2161959730 @default.
- W2296394872 cites W2294059674 @default.
- W2296394872 cites W2299971393 @default.
- W2296394872 cites W2800002835 @default.
- W2296394872 cites W2963403336 @default.
- W2296394872 cites W3101788651 @default.
- W2296394872 cites W610284107 @default.
- W2296394872 doi "https://doi.org/10.1109/icmla.2015.109" @default.
- W2296394872 hasPublicationYear "2015" @default.
- W2296394872 type Work @default.
- W2296394872 sameAs 2296394872 @default.
- W2296394872 citedByCount "6" @default.
- W2296394872 countsByYear W22963948722015 @default.
- W2296394872 countsByYear W22963948722016 @default.
- W2296394872 countsByYear W22963948722017 @default.
- W2296394872 countsByYear W22963948722018 @default.
- W2296394872 crossrefType "proceedings-article" @default.
- W2296394872 hasAuthorship W2296394872A5025477216 @default.
- W2296394872 hasAuthorship W2296394872A5035780973 @default.
- W2296394872 hasAuthorship W2296394872A5081192463 @default.
- W2296394872 hasConcept C105795698 @default.
- W2296394872 hasConcept C11413529 @default.
- W2296394872 hasConcept C126255220 @default.
- W2296394872 hasConcept C138885662 @default.
- W2296394872 hasConcept C154945302 @default.
- W2296394872 hasConcept C178650346 @default.
- W2296394872 hasConcept C180877172 @default.
- W2296394872 hasConcept C185142706 @default.
- W2296394872 hasConcept C185429906 @default.
- W2296394872 hasConcept C1921717 @default.
- W2296394872 hasConcept C33923547 @default.
- W2296394872 hasConcept C41008148 @default.
- W2296394872 hasConcept C41895202 @default.
- W2296394872 hasConcept C7149132 @default.
- W2296394872 hasConcept C73555534 @default.
- W2296394872 hasConceptScore W2296394872C105795698 @default.
- W2296394872 hasConceptScore W2296394872C11413529 @default.
- W2296394872 hasConceptScore W2296394872C126255220 @default.
- W2296394872 hasConceptScore W2296394872C138885662 @default.
- W2296394872 hasConceptScore W2296394872C154945302 @default.
- W2296394872 hasConceptScore W2296394872C178650346 @default.
- W2296394872 hasConceptScore W2296394872C180877172 @default.
- W2296394872 hasConceptScore W2296394872C185142706 @default.
- W2296394872 hasConceptScore W2296394872C185429906 @default.
- W2296394872 hasConceptScore W2296394872C1921717 @default.
- W2296394872 hasConceptScore W2296394872C33923547 @default.
- W2296394872 hasConceptScore W2296394872C41008148 @default.
- W2296394872 hasConceptScore W2296394872C41895202 @default.
- W2296394872 hasConceptScore W2296394872C7149132 @default.
- W2296394872 hasConceptScore W2296394872C73555534 @default.
- W2296394872 hasLocation W22963948721 @default.
- W2296394872 hasOpenAccess W2296394872 @default.
- W2296394872 hasPrimaryLocation W22963948721 @default.
- W2296394872 hasRelatedWork W1910942334 @default.
- W2296394872 hasRelatedWork W2029795429 @default.
- W2296394872 hasRelatedWork W2070774586 @default.
- W2296394872 hasRelatedWork W2138030803 @default.
- W2296394872 hasRelatedWork W2181933175 @default.
- W2296394872 hasRelatedWork W3009379717 @default.
- W2296394872 hasRelatedWork W3023135770 @default.
- W2296394872 hasRelatedWork W3124202622 @default.
- W2296394872 hasRelatedWork W3159295116 @default.
- W2296394872 hasRelatedWork W4255757349 @default.
- W2296394872 isParatext "false" @default.
- W2296394872 isRetracted "false" @default.
- W2296394872 magId "2296394872" @default.
- W2296394872 workType "article" @default.