Matches in SemOpenAlex for { <https://semopenalex.org/work/W3035721323> ?p ?o ?g. }
Showing items 1 to 100 of
100
with 100 items per page.
- W3035721323 abstract "Computer models are typically used to describe the input-output relationships in physical systems. These models require a set of known inputs besides parameters that need to be specified. Often, these unknown parameters are calibrated using a limited amount of field data. The seminal paper by Kennedy and O'Hagan (2001) introduced a statistical approach based on the Gaussian process model that allows for quantification of uncertainties due to: a) model parameters (calibration), b) limited number of simulations, c) discrepancy between the simulation code and the actual physical system, and d) observation process. The approach is very robust and has been widely applied to cases where physical systems are described by computationally expensive simulations. In this paper, we propose a framework for bias estimation and calibration of computer models using graph theory and neural networks. Similarly to the Kennedy and O'Hagan approach, we use observed data to simultaneously calibrate model parameters and estimate the discrepancy term through a metamodel. While Kennedy and O'Hagan (2001) based their approach on the Gaussian process model, we choose neural networks. Another difference is that we represent the hybrid model that merges the original computer model with the neural network through a graph. This allows us to easily estimate the discrepancy term even for hidden nodes of the graph. We studied the performance of our framework with the aid of numerical experiments and state-of-the-art machine learning software packages. The preliminary results are promising and demonstrate the ability to perform simultaneous estimation of discrepancy and model parameters at reasonable computational cost." @default.
- W3035721323 created "2020-06-19" @default.
- W3035721323 creator A5000968828 @default.
- W3035721323 creator A5001649611 @default.
- W3035721323 creator A5003577193 @default.
- W3035721323 creator A5062425983 @default.
- W3035721323 date "2020-06-08" @default.
- W3035721323 modified "2023-10-04" @default.
- W3035721323 title "Estimating Parameters and Discrepancy of Computer Models with Graphs and Neural Networks" @default.
- W3035721323 cites W1973333099 @default.
- W3035721323 cites W1982886636 @default.
- W3035721323 cites W2002016471 @default.
- W3035721323 cites W2012815280 @default.
- W3035721323 cites W2066724746 @default.
- W3035721323 cites W2072028796 @default.
- W3035721323 cites W2078454401 @default.
- W3035721323 cites W2080556931 @default.
- W3035721323 cites W2092612581 @default.
- W3035721323 cites W2130416410 @default.
- W3035721323 cites W2158518844 @default.
- W3035721323 cites W2180650559 @default.
- W3035721323 cites W2226462370 @default.
- W3035721323 cites W2286036836 @default.
- W3035721323 cites W2761434131 @default.
- W3035721323 cites W2885397520 @default.
- W3035721323 cites W2985735500 @default.
- W3035721323 cites W3153089761 @default.
- W3035721323 doi "https://doi.org/10.2514/6.2020-3123" @default.
- W3035721323 hasPublicationYear "2020" @default.
- W3035721323 type Work @default.
- W3035721323 sameAs 3035721323 @default.
- W3035721323 citedByCount "3" @default.
- W3035721323 countsByYear W30357213232021 @default.
- W3035721323 countsByYear W30357213232022 @default.
- W3035721323 crossrefType "proceedings-article" @default.
- W3035721323 hasAuthorship W3035721323A5000968828 @default.
- W3035721323 hasAuthorship W3035721323A5001649611 @default.
- W3035721323 hasAuthorship W3035721323A5003577193 @default.
- W3035721323 hasAuthorship W3035721323A5062425983 @default.
- W3035721323 hasConcept C105795698 @default.
- W3035721323 hasConcept C11413529 @default.
- W3035721323 hasConcept C116672817 @default.
- W3035721323 hasConcept C119857082 @default.
- W3035721323 hasConcept C121332964 @default.
- W3035721323 hasConcept C132525143 @default.
- W3035721323 hasConcept C154945302 @default.
- W3035721323 hasConcept C163716315 @default.
- W3035721323 hasConcept C165838908 @default.
- W3035721323 hasConcept C177264268 @default.
- W3035721323 hasConcept C199360897 @default.
- W3035721323 hasConcept C2777904410 @default.
- W3035721323 hasConcept C33923547 @default.
- W3035721323 hasConcept C41008148 @default.
- W3035721323 hasConcept C43126263 @default.
- W3035721323 hasConcept C50644808 @default.
- W3035721323 hasConcept C61326573 @default.
- W3035721323 hasConcept C62520636 @default.
- W3035721323 hasConcept C80444323 @default.
- W3035721323 hasConcept C81692654 @default.
- W3035721323 hasConcept C86610423 @default.
- W3035721323 hasConcept C98045186 @default.
- W3035721323 hasConceptScore W3035721323C105795698 @default.
- W3035721323 hasConceptScore W3035721323C11413529 @default.
- W3035721323 hasConceptScore W3035721323C116672817 @default.
- W3035721323 hasConceptScore W3035721323C119857082 @default.
- W3035721323 hasConceptScore W3035721323C121332964 @default.
- W3035721323 hasConceptScore W3035721323C132525143 @default.
- W3035721323 hasConceptScore W3035721323C154945302 @default.
- W3035721323 hasConceptScore W3035721323C163716315 @default.
- W3035721323 hasConceptScore W3035721323C165838908 @default.
- W3035721323 hasConceptScore W3035721323C177264268 @default.
- W3035721323 hasConceptScore W3035721323C199360897 @default.
- W3035721323 hasConceptScore W3035721323C2777904410 @default.
- W3035721323 hasConceptScore W3035721323C33923547 @default.
- W3035721323 hasConceptScore W3035721323C41008148 @default.
- W3035721323 hasConceptScore W3035721323C43126263 @default.
- W3035721323 hasConceptScore W3035721323C50644808 @default.
- W3035721323 hasConceptScore W3035721323C61326573 @default.
- W3035721323 hasConceptScore W3035721323C62520636 @default.
- W3035721323 hasConceptScore W3035721323C80444323 @default.
- W3035721323 hasConceptScore W3035721323C81692654 @default.
- W3035721323 hasConceptScore W3035721323C86610423 @default.
- W3035721323 hasConceptScore W3035721323C98045186 @default.
- W3035721323 hasLocation W30357213231 @default.
- W3035721323 hasOpenAccess W3035721323 @default.
- W3035721323 hasPrimaryLocation W30357213231 @default.
- W3035721323 hasRelatedWork W1635518421 @default.
- W3035721323 hasRelatedWork W1968523686 @default.
- W3035721323 hasRelatedWork W2063381173 @default.
- W3035721323 hasRelatedWork W2318821300 @default.
- W3035721323 hasRelatedWork W2783363666 @default.
- W3035721323 hasRelatedWork W3088831819 @default.
- W3035721323 hasRelatedWork W3124980497 @default.
- W3035721323 hasRelatedWork W3141541034 @default.
- W3035721323 hasRelatedWork W4213436830 @default.
- W3035721323 hasRelatedWork W4245715244 @default.
- W3035721323 isParatext "false" @default.
- W3035721323 isRetracted "false" @default.
- W3035721323 magId "3035721323" @default.
- W3035721323 workType "article" @default.