Deep Gaussian Process metamodeling of sequentially sampled non-stationary response surfaces

Dutordoir, Vincent; Knudde, Nicolas; Herten, Joachim van der; Couckuyt, Ivo; Dhaene, Tom

doi:10.1109/wsc.2017.8247911

Cited by 6 publications

(7 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance measures are calculated using the formulae shown in numbers 6 through 9. Also, the loss function 58 was used as a performance metric.…”

Section: Model Evaluationmentioning

confidence: 99%

“…Deep Gaussian processes are now used to solve the drawbacks of Gaussian process models. 15 Iteratively convolving several GP functions across an image is accomplished using a deep convolutional Gaussian process. 16 Traditional (single-layer) GPs have limitations, and a non-parametric deep Bayesian method known as a deep Gaussian process (DGP) uses a hierarchical composition of GPs to address these limitations while maintaining the benefits of GPs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Classification of cassava leaf diseases using deep Gaussian transfer learning model

Ahishakiye

Mwangi

Murithi

et al. 2023

Engineering Reports

View full text Add to dashboard Cite

In Sub‐Saharan Africa, experts visually examine the plants and look for disease symptoms on the leaves to diagnose cassava diseases, a subjective method. Machine learning algorithms have been employed to quickly identify and classify crop diseases. In this study, we propose a model that integrates a transfer learning approach with a deep Gaussian convolutional neural network model. In this study, two pre‐trained transfer learning models were used, that is, MobileNet V2 and VGG16, together with three different kernels: a hybrid kernel (a product of a squared exponential kernel and a rational quadratic kernel), a squared exponential kernel, and a rational quadratic kernel. In experiments using MobileNet V2 and the three kernels, the hybrid kernel performed better, with an accuracy of 90.11%, compared to 86.03% and 85.14% for the squared exponential kernel and a rational quadratic kernel, respectively. Additionally, experiments using VGG16 and the three kernels showed that the hybrid kernel performed better, with an accuracy of 88.63%, compared to the squared exponential kernel's accuracy of 84.62% and the rational quadratic kernel's accuracy of 83.95%, respectively. All the experiments were done using a traditional computer with no access to GPU and this was the major limitation of the study.

show abstract

“…The performance measures are calculated using the formulae shown in numbers 6 through 9. Also, the loss function 58 was used as a performance metric.…”

Section: Model Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Classification of cassava leaf diseases using deep Gaussian transfer learning model

Ahishakiye

Mwangi

Murithi

et al. 2023

Engineering Reports

View full text Add to dashboard Cite

show abstract

“…Deep Gaussian Process Regression (DGPR) is evolved on the basis of GPR which has been widely used due to its analytical tractability, a training procedure which automatically guards against overfitting and ability to provide a full predictive distribution over the outputs [22]. In detail, the GPR implementation can be realized by an analytic solution, but the selection of a single kernel function may restrict the prediction ability of the model.…”

Section: Deep Gaussian Process Regressionmentioning

confidence: 99%

“…Different from the single-layer GPR which can only represent a restricted class of functions, DGPR is a statistical multi-layer hierarchical model of GPR [23] which the output of the current layer acts as the next layer input and enables a deep probabilistic nonparametric approach to flexibly tackle complex machine-learning problems with sound quantification of uncertainty [20]. In other words, DGPR retains useful properties of GPR [22] and overcomes the limitations of GPR, which is suited for representation of highly complex data relationships.…”

Section: Deep Gaussian Process Regressionmentioning

confidence: 99%

Deep Gaussian Process Regression for Performance Improvement of POS During GPS Outages

et al. 2020

View full text Add to dashboard Cite

Position and orientation system (POS) is a high-precision inertial navigation systems /Global navigation satellite system (INS/GNSS) integrated system that can continuously provide timespatial reference for airborne remote sensing, mobile mapping and vehicle localization using Kalman Filter (KF) during the availability of GPS measurements. However, the POS suffers from the GPS outages in some especial environment, whose accuracy of motion parameters degrades with the time accumulation. In order to suppress the performance degradation caused by GPS outages, this work proposes a hybrid predictor based on data-driven Deep Gaussian Process Regression (DGPR), which uses multi-layer Gaussian Process Regression to deal with highly complex data relationships and predictive uncertainty. Once GPS outages happen, the proposed approach starts to predict the observation measurement, and then feeds it to KF as a virtual update to estimate all the INS errors. The proposed approach is validated by the real flight test, and the experimental results show that significant performance improvement has been achieved during various GPS outages.

show abstract

“…Some inroads have been made along these lines. Dutordoir et al (2017) fit DGPs to sequentially collected data, but acquisition criteria were not based on the DGP fits. Rajaram et al (2020) suggested a maximum variance criterion, a strategy sometimes called "active learning MacKay" (ALM; MacKay, 1992), with DGPs.…”

Section: Introductionmentioning

confidence: 99%

Active Learning for Deep Gaussian Process Surrogates

Sauer¹,

Gramacy²,

Higdon³

2020

Preprint

View full text Add to dashboard Cite

Deep Gaussian processes (DGPs) are increasingly popular as predictive models in machine learning (ML) for their non-stationary flexibility and ability to cope with abrupt regime changes in training data. Here we explore DGPs as surrogates for computer simulation experiments whose response surfaces exhibit similar characteristics. In particular, we transport a DGP's automatic warping of the input space and full uncertainty quantification (UQ), via a novel elliptical slice sampling (ESS) Bayesian posterior inferential scheme, through to active learning (AL) strategies that distribute runs non-uniformly in the input space -something an ordinary (stationary) GP could not do. Building up the design sequentially in this way allows smaller training sets, limiting both expensive evaluation of the simulator code and mitigating cubic costs of DGP inference. When training data sizes are kept small through careful acquisition, and with parsimonious layout of latent layers, the framework can be both effective and computationally tractable. Our methods are illustrated on simulation data and two real computer experiments of varying input dimensionality. We provide an open source implementation in the deepgp package on CRAN.

show abstract

Deep Gaussian Process metamodeling of sequentially sampled non-stationary response surfaces

Cited by 6 publications

References 15 publications

Classification of cassava leaf diseases using deep Gaussian transfer learning model

Classification of cassava leaf diseases using deep Gaussian transfer learning model

Deep Gaussian Process Regression for Performance Improvement of POS During GPS Outages

Active Learning for Deep Gaussian Process Surrogates

Contact Info

Product

Resources

About