How to fit a properly nonlinear classification model from conventional well logs to lithofacies is a key problem for machine learning methods.Kernel methods(e.g.,KFD,SVM,MSVM)are effective attempts to solve this issue...How to fit a properly nonlinear classification model from conventional well logs to lithofacies is a key problem for machine learning methods.Kernel methods(e.g.,KFD,SVM,MSVM)are effective attempts to solve this issue due to abilities of handling nonlinear features by kernel functions.Deep mining of log features indicating lithofacies still needs to be improved for kernel methods.Hence,this work employs deep neural networks to enhance the kernel principal component analysis(KPCA)method and proposes a deep kernel method(DKM)for lithofacies identification using well logs.DKM includes a feature extractor and a classifier.The feature extractor consists of a series of KPCA models arranged according to residual network structure.A gradient-free optimization method is introduced to automatically optimize parameters and structure in DKM,which can avoid complex tuning of parameters in models.To test the validation of the proposed DKM for lithofacies identification,an open-sourced dataset with seven con-ventional logs(GR,CAL,AC,DEN,CNL,LLD,and LLS)and lithofacies labels from the Daniudi Gas Field in China is used.There are eight lithofacies,namely clastic rocks(pebbly,coarse,medium,and fine sand-stone,siltstone,mudstone),coal,and carbonate rocks.The comparisons between DKM and three commonly used kernel methods(KFD,SVM,MSVM)show that(1)DKM(85.7%)outperforms SVM(77%),KFD(79.5%),and MSVM(82.8%)in accuracy of lithofacies identification;(2)DKM is about twice faster than the multi-kernel method(MSVM)with good accuracy.The blind well test in Well D13 indicates that compared with the other three methods DKM improves about 24%in accuracy,35%in precision,41%in recall,and 40%in F1 score,respectively.In general,DKM is an effective method for complex lithofacies identification.This work also discussed the optimal structure and classifier for DKM.Experimental re-sults show that(m_(1),m_(2),O)is the optimal model structure and linear svM is the optimal classifier.(m_(1),m_(2),O)means there are m KPCAs,and then m2 residual units.A workflow to determine an optimal classifier in DKM for lithofacies identification is proposed,too.展开更多
We report the virtual instrumentation of both time-domain(TD)and spectral-domain(SD)opticai coherence tomography(OCT)systems.With a virtual partial coherence source fromeither a simulated or measured spectrum,the OCT ...We report the virtual instrumentation of both time-domain(TD)and spectral-domain(SD)opticai coherence tomography(OCT)systems.With a virtual partial coherence source fromeither a simulated or measured spectrum,the OCT signals of both A-scan and B-scan weredemonstrated.The spectrometric detector's pixel number,dynamic range,noise,as well asspectral resolution can be simulated in the virtual spectral domain(SD-OCT).The virtual-OCT system provides an environment for parameter evaluation and algorithm optimization for ex-perimental OCT instrumentation,and promotes the understanding of OCT imaging and signal post-processing processes.The authors thank Dr.Thomas FitzGibbon forcomments on earlier drafts of the manuscript.展开更多
Super-resolution structured ilumination microscopy(SR-SIM)is finding increasing application in biomedical research due to its superior ability to visualize subcellular dynamics in living cells.However,during image rec...Super-resolution structured ilumination microscopy(SR-SIM)is finding increasing application in biomedical research due to its superior ability to visualize subcellular dynamics in living cells.However,during image reconstruction artifacts can be introduced and when coupled with time-consuming postprocessing procedures,limits this technique from becoming a routine imaging tool for biologists.To address these issues,an accelerated,artifact-reduced reconstruction algorithm termed joint space frequency reconstruction-based artifact reduction algorithm(JSFR-AR-SIM)was developed by integrating a high-speed reconstruc tion framework with a high-fidelity optimization approach designed to suppress the sidelobe artifact.Consequently,JSFR-AR-SIM produces high-quality,super-resolution images with minimal artifacts,and the reconstruction speed is increased.We anticipate this algorithm to facilitate SR-SIM becoming a routine tool in biomedical laboratories.展开更多
The establishment of a possible connection between neutrino emission and gravitational-wave(GW)bursts is important to our understanding of the physical processes that occur when black holes or neutron stars merge.In t...The establishment of a possible connection between neutrino emission and gravitational-wave(GW)bursts is important to our understanding of the physical processes that occur when black holes or neutron stars merge.In the Daya Bay experiment,using the data collected from December 2011 to August 2017,a search was per-formed for electron-antineutrino signals that coincided with detected GW events,including GW150914,GW151012,GW151226,GW170104,GW170608,GW 170814,and GW 170817.We used three time windows of±10,±500,and±1000 s relative to the occurrence of the GW events and a neutrino energy range of 1.8 to 100 MeV to search for correlated neutrino candidates.The detected electron-antineutrino candidates were consistent with the expected background rates for all the three time windows.Assuming monochromatic spectra,we found upper limits(90%confidence level)of the electron-antineutrino fluence of(1.13-2.44)×10^(11)cm^(-2)at 5 MeV to 8.0×10^(7)cm^(-2)at 100 MeV for the three time w indows.Under the assumption of a Fermi-Dirac spectrum,the upper limits were found to be(5.4-7.0)×10^(9)cm^(2)for the three time windows.展开更多
基金supported by the National Natural Science Foundation of China(Grant No.42002134)China Postdoctoral Science Foundation(Grant No.2021T140735)Science Foundation of China University of Petroleum,Beijing(Grant Nos.2462020XKJS02 and 2462020YXZZ004).
文摘How to fit a properly nonlinear classification model from conventional well logs to lithofacies is a key problem for machine learning methods.Kernel methods(e.g.,KFD,SVM,MSVM)are effective attempts to solve this issue due to abilities of handling nonlinear features by kernel functions.Deep mining of log features indicating lithofacies still needs to be improved for kernel methods.Hence,this work employs deep neural networks to enhance the kernel principal component analysis(KPCA)method and proposes a deep kernel method(DKM)for lithofacies identification using well logs.DKM includes a feature extractor and a classifier.The feature extractor consists of a series of KPCA models arranged according to residual network structure.A gradient-free optimization method is introduced to automatically optimize parameters and structure in DKM,which can avoid complex tuning of parameters in models.To test the validation of the proposed DKM for lithofacies identification,an open-sourced dataset with seven con-ventional logs(GR,CAL,AC,DEN,CNL,LLD,and LLS)and lithofacies labels from the Daniudi Gas Field in China is used.There are eight lithofacies,namely clastic rocks(pebbly,coarse,medium,and fine sand-stone,siltstone,mudstone),coal,and carbonate rocks.The comparisons between DKM and three commonly used kernel methods(KFD,SVM,MSVM)show that(1)DKM(85.7%)outperforms SVM(77%),KFD(79.5%),and MSVM(82.8%)in accuracy of lithofacies identification;(2)DKM is about twice faster than the multi-kernel method(MSVM)with good accuracy.The blind well test in Well D13 indicates that compared with the other three methods DKM improves about 24%in accuracy,35%in precision,41%in recall,and 40%in F1 score,respectively.In general,DKM is an effective method for complex lithofacies identification.This work also discussed the optimal structure and classifier for DKM.Experimental re-sults show that(m_(1),m_(2),O)is the optimal model structure and linear svM is the optimal classifier.(m_(1),m_(2),O)means there are m KPCAs,and then m2 residual units.A workflow to determine an optimal classifier in DKM for lithofacies identification is proposed,too.
基金supported by the National Instrumentation Program(2013YQ03065102)the "973"Major State Basic Research Development Program of China(2011CB707502,2010CB933901,2011CB809101)the National Natural Science Foundation of China(61178076,61307015).
文摘We report the virtual instrumentation of both time-domain(TD)and spectral-domain(SD)opticai coherence tomography(OCT)systems.With a virtual partial coherence source fromeither a simulated or measured spectrum,the OCT signals of both A-scan and B-scan weredemonstrated.The spectrometric detector's pixel number,dynamic range,noise,as well asspectral resolution can be simulated in the virtual spectral domain(SD-OCT).The virtual-OCT system provides an environment for parameter evaluation and algorithm optimization for ex-perimental OCT instrumentation,and promotes the understanding of OCT imaging and signal post-processing processes.The authors thank Dr.Thomas FitzGibbon forcomments on earlier drafts of the manuscript.
基金supported by the National Key Research and Development Program of China(2022YFF0712500)the Natural Science Foundation of China(NSFC)(62135003,62005208,62205267,12204380)+3 种基金the Innovation Capability Support Program of Shaanxi(program no.2021TD-57)the Natural Science Basic Research Program of Shaanxi(2022JZ-34,2020JQ-072,2022JQ-069)NIH grants GM144414 to P.R.B.a Preliminary Data and Application Preparation Grant to P.R.B.and K.W.We appreciate Standard Imaging(Beijing)Biotechnology Co.Ltd for assistance with sample preparation.
文摘Super-resolution structured ilumination microscopy(SR-SIM)is finding increasing application in biomedical research due to its superior ability to visualize subcellular dynamics in living cells.However,during image reconstruction artifacts can be introduced and when coupled with time-consuming postprocessing procedures,limits this technique from becoming a routine imaging tool for biologists.To address these issues,an accelerated,artifact-reduced reconstruction algorithm termed joint space frequency reconstruction-based artifact reduction algorithm(JSFR-AR-SIM)was developed by integrating a high-speed reconstruc tion framework with a high-fidelity optimization approach designed to suppress the sidelobe artifact.Consequently,JSFR-AR-SIM produces high-quality,super-resolution images with minimal artifacts,and the reconstruction speed is increased.We anticipate this algorithm to facilitate SR-SIM becoming a routine tool in biomedical laboratories.
基金Daya Bay is supported in part by the Ministry of Science and Technology o f China, the U.S. Department o f Energy, the Chinese Academy of Sciences, the CASCenter for Excellence in Particle Physics, the National Natural Science Foundation of China, the Guangdong provincial government, the Shenzhen municipal government,the China General Nuclear Power Group, Key Laboratory of Particle and Radiation Imaging (Tsinghua University), the Ministry of Education, Key Laboratory ofParticle Physics and Particle Irradiation (Shandong University), the Ministry o f Education, Shanghai Laboratory for Particle Physics and Cosmology, the ResearchGrants Council o f the Hong Kong Special Administrative Region of China, the University Development Fund of the University of Hong Kong, the MOE program forResearch of Excellence at National Taiwan University, National Chiao-Tung University, NSC fund support from Taiwan, the U.S. National Science Foundation, the AlfredP. Sloan Foundation, the Ministry o f Education, Youth, and Sports of the Czech Republic, the Charles University GAUK (284317), the Joint Institute o f NuclearResearch in Dubna, Russia, the National Commission of Scientific and Technological Research of Chile, and the Tsinghua University Initiative Scientific Research Program.
文摘The establishment of a possible connection between neutrino emission and gravitational-wave(GW)bursts is important to our understanding of the physical processes that occur when black holes or neutron stars merge.In the Daya Bay experiment,using the data collected from December 2011 to August 2017,a search was per-formed for electron-antineutrino signals that coincided with detected GW events,including GW150914,GW151012,GW151226,GW170104,GW170608,GW 170814,and GW 170817.We used three time windows of±10,±500,and±1000 s relative to the occurrence of the GW events and a neutrino energy range of 1.8 to 100 MeV to search for correlated neutrino candidates.The detected electron-antineutrino candidates were consistent with the expected background rates for all the three time windows.Assuming monochromatic spectra,we found upper limits(90%confidence level)of the electron-antineutrino fluence of(1.13-2.44)×10^(11)cm^(-2)at 5 MeV to 8.0×10^(7)cm^(-2)at 100 MeV for the three time w indows.Under the assumption of a Fermi-Dirac spectrum,the upper limits were found to be(5.4-7.0)×10^(9)cm^(2)for the three time windows.