Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037133998> ?p ?o ?g. }
- W3037133998 abstract "Recurrent neural networks (RNNs) are instrumental in modelling sequential and time-series data. Yet, when using RNNs to inform decision-making, predictions by themselves are not sufficient; we also need estimates of predictive uncertainty. Existing approaches for uncertainty quantification in RNNs are based predominantly on Bayesian methods; these are computationally prohibitive, and require major alterations to the RNN architecture and training. Capitalizing on ideas from classical jackknife resampling, we develop a frequentist alternative that: (a) does not interfere with model training or compromise its accuracy, (b) applies to any RNN architecture, and (c) provides theoretical coverage guarantees on the estimated uncertainty intervals. Our method derives predictive uncertainty from the variability of the (jackknife) sampling distribution of the RNN outputs, which is estimated by repeatedly deleting blocks of (temporally-correlated) training data, and collecting the predictions of the RNN re-trained on the remaining data. To avoid exhaustive re-training, we utilize influence functions to estimate the effect of removing training data blocks on the learned RNN parameters. Using data from a critical care setting, we demonstrate the utility of uncertainty quantification in sequential decision-making." @default.
- W3037133998 created "2020-07-02" @default.
- W3037133998 creator A5012339002 @default.
- W3037133998 creator A5014596965 @default.
- W3037133998 date "2020-06-20" @default.
- W3037133998 modified "2023-10-17" @default.
- W3037133998 title "Frequentist Uncertainty in Recurrent Neural Networks via Blockwise Influence Functions" @default.
- W3037133998 cites W1568278035 @default.
- W3037133998 cites W16706201 @default.
- W3037133998 cites W1719489212 @default.
- W3037133998 cites W1753482797 @default.
- W3037133998 cites W1911155832 @default.
- W3037133998 cites W1981492171 @default.
- W3037133998 cites W2000291746 @default.
- W3037133998 cites W2006903949 @default.
- W3037133998 cites W2012712694 @default.
- W3037133998 cites W2047694274 @default.
- W3037133998 cites W2076237237 @default.
- W3037133998 cites W2096533821 @default.
- W3037133998 cites W2102890221 @default.
- W3037133998 cites W2127141656 @default.
- W3037133998 cites W2130942839 @default.
- W3037133998 cites W2136658108 @default.
- W3037133998 cites W2144578442 @default.
- W3037133998 cites W2160597133 @default.
- W3037133998 cites W2167433878 @default.
- W3037133998 cites W2172198266 @default.
- W3037133998 cites W2207674461 @default.
- W3037133998 cites W2208550830 @default.
- W3037133998 cites W2216680303 @default.
- W3037133998 cites W2242047860 @default.
- W3037133998 cites W2271982975 @default.
- W3037133998 cites W2396881363 @default.
- W3037133998 cites W2402268235 @default.
- W3037133998 cites W2583977184 @default.
- W3037133998 cites W2597603852 @default.
- W3037133998 cites W2606554264 @default.
- W3037133998 cites W2734777338 @default.
- W3037133998 cites W2751802138 @default.
- W3037133998 cites W2767630563 @default.
- W3037133998 cites W2773625660 @default.
- W3037133998 cites W2774513877 @default.
- W3037133998 cites W2786857698 @default.
- W3037133998 cites W2883695301 @default.
- W3037133998 cites W2889928394 @default.
- W3037133998 cites W2892273226 @default.
- W3037133998 cites W2912168444 @default.
- W3037133998 cites W2921829327 @default.
- W3037133998 cites W2963215553 @default.
- W3037133998 cites W2963238274 @default.
- W3037133998 cites W2963266340 @default.
- W3037133998 cites W2963561234 @default.
- W3037133998 cites W2963842982 @default.
- W3037133998 cites W2964059111 @default.
- W3037133998 cites W2964201648 @default.
- W3037133998 cites W2964232608 @default.
- W3037133998 cites W2964872595 @default.
- W3037133998 cites W2970311848 @default.
- W3037133998 cites W2970631161 @default.
- W3037133998 cites W2970859221 @default.
- W3037133998 cites W2971113110 @default.
- W3037133998 cites W3126856044 @default.
- W3037133998 cites W2108719358 @default.
- W3037133998 hasPublicationYear "2020" @default.
- W3037133998 type Work @default.
- W3037133998 sameAs 3037133998 @default.
- W3037133998 citedByCount "2" @default.
- W3037133998 countsByYear W30371339982020 @default.
- W3037133998 countsByYear W30371339982021 @default.
- W3037133998 crossrefType "posted-content" @default.
- W3037133998 hasAuthorship W3037133998A5012339002 @default.
- W3037133998 hasAuthorship W3037133998A5014596965 @default.
- W3037133998 hasConcept C105795698 @default.
- W3037133998 hasConcept C107673813 @default.
- W3037133998 hasConcept C119857082 @default.
- W3037133998 hasConcept C147168706 @default.
- W3037133998 hasConcept C150921843 @default.
- W3037133998 hasConcept C154945302 @default.
- W3037133998 hasConcept C160234255 @default.
- W3037133998 hasConcept C162376815 @default.
- W3037133998 hasConcept C185429906 @default.
- W3037133998 hasConcept C33923547 @default.
- W3037133998 hasConcept C41008148 @default.
- W3037133998 hasConcept C50644808 @default.
- W3037133998 hasConcept C81790035 @default.
- W3037133998 hasConceptScore W3037133998C105795698 @default.
- W3037133998 hasConceptScore W3037133998C107673813 @default.
- W3037133998 hasConceptScore W3037133998C119857082 @default.
- W3037133998 hasConceptScore W3037133998C147168706 @default.
- W3037133998 hasConceptScore W3037133998C150921843 @default.
- W3037133998 hasConceptScore W3037133998C154945302 @default.
- W3037133998 hasConceptScore W3037133998C160234255 @default.
- W3037133998 hasConceptScore W3037133998C162376815 @default.
- W3037133998 hasConceptScore W3037133998C185429906 @default.
- W3037133998 hasConceptScore W3037133998C33923547 @default.
- W3037133998 hasConceptScore W3037133998C41008148 @default.
- W3037133998 hasConceptScore W3037133998C50644808 @default.
- W3037133998 hasConceptScore W3037133998C81790035 @default.
- W3037133998 hasLocation W30371339981 @default.
- W3037133998 hasOpenAccess W3037133998 @default.