检索结果-南通市图书馆

Conditional Coverage Estimation for High-Quality Prediction Intervals

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems Science and Systems Engineering 2023年第3期32卷 289-319页

作者： Ziyi Huang Henry lam Haofeng Zhang School of Engineering and Applied Science Columbia UniversityNew YorkNY 10027USA

Deep learning has been recently studied to generate high-quality prediction intervals(PIs)for uncertainty quantification in regression tasks,including recent applications in simulation *** high-quality criterion requires PIs to be as narrow as possible,whilst maintaining a pre-specified level of data(marginal)***,most existing works for high-quality PIs lack accurate information on conditional coverage,which may cause unreliable predictions if it is significantly smaller than the marginal *** address this problem,we propose an end-to-end framework which could output high-quality PIs and simultaneously provide their conditional coverage *** doing so,we design a new loss function that is both easy-to-implement and theoretically justified via an exponential concentration *** evaluation on real-world benchmark datasets and synthetic examples shows that our approach not only achieves competitive results on high-quality PIs in terms of average PI width,but also accurately estimates conditional coverage information that is useful in assessing model uncertainty.

关键词： Uncertainty quantification prediction intervals conditional coverage neural networks calibrationerror

An unsupervised learning approach to study synchroneity of past events in the South China Sea

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Earth Science 2019年第3期13卷 628-640页

作者： Kevin C. Tse Hon-Chim Chiu Man-yin Tsang yiliang Li Edmund y. lam Department of Earth Sciences The University of Hong KongPokfulamHong KongChina Department of Geography and Centre for Geo-computation Studies Hong Kong Baptist UniversityKowloon TongHong KongChina Department of Earth Sciences University of TorontoCanada Department of Electrical and Electronic Engineering The University of Hong KongPokfulamHong KongChina

Unsupervised machine learning methods were applied on multivariate geophysical and geochemical datasets of ocean floor sediment cores collected from the South China Sea. The well-preserved and continuous core samples comprising high resolution Cenozoic sediment records enable scientists to carry out paleoenvironment studies in detail. Bayesian age-depth chronological models constructed from biostratigraphic control points for the drilling sites are applied on cluster boundaries generated from two popular unsupervised learning methods: K-means and random forest. The unsupervised learning methods experimented have produced compact and unambiguous clusters from the datasets, indicating that previously unknown data patterns can be revealed when all variables from the datasets are taken into account simultaneously. A study of synchroneity of past events represented by the cluster boundaries across geographically separated ocean drilling sites is achieved through converting the fixed depths of cluster boundaries into chronological ranges represented by Gaussian density plots which are then compared with known past events in the region. A Gaussian density peak at around 7.2 Ma has been identified from results of all three sites and it is suggested to coincide with the initiation of the East Asian monsoon. Contrary to traditional statistical approach, a priori assumptions are not required for unsupervised learning, and the clustering results serve as a novel data-driven proxy for studying the complex and dynamic processes of the paleoenvironment surrounding the ocean sediment. This work serves as a pioneering approach to extract valuable information of regional events and opens up a systematic and objective way to study the vast global ocean sediment datasets.

关键词： machine learning ocean sediments unsupervised classification

Unsupervised learning on scientific ocean drilling datasets from the South China Sea

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Earth Science 2019年第1期13卷 180-190页

作者： Kevin C.TSE Hon-Chim CHIU Man-yin TSANG yiliang LI Edmund y.lam Department of Earth Sciences The University of Hong KongPokfulamHong KongChina Department of Geography and Centre for Geo-computation Studies Hong Kong Baptist UniversityKowloon TongHong KongChina Department of Earth Sciences University of TorontoTorontoON M5S 2M8Canada Department of Electrical and Electronic Engineering The University of Hong KongPokfulamHong KongChina

Unsupervised learning methods were applied to explore data patterns in multivariate geophysical datasets collected from ocean floor sediment core samples coming from scientific ocean drilling in the South China *** to studies on similar datasets,but using supervised learning methods which are designed to make predictions based on sample training data,unsupervised learning methods require no a priori information and focus only on the input *** this study,popular unsupervised learning methods including K-means,self-organizing maps,hierarchical clustering and random forest were coupled with different distance metrics to form exploratory data *** resulting data clusters were externally validated with lithologic units and geologic time scales assigned to the datasets by conventional methods. Compact and connected data clusters displayed varying degrees of correspondence with existing classification by lithologic units and geologic time scales.K-means and self-organizing maps were observed to perform better with lithologic units while random forest corresponded best with geologic time *** study sets a pioneering example of how unsupervised machine learning methods can be used as an automatic processing tool for the increasingly high volume of scientific ocean drilling data.

关键词： machine learning unsupervised learning ODP IODP clustering

Complete characterization of ultrashort optical pulses with a phase-shifting wedged reversal shearing interferometer

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Light(Science & Applications) 2018年第1期7卷 755-765页

作者： Billy lam Chunlei Guo The Institute of Optics University of RochesterRochesterNY 14627USA Changchun Institute of Optics Fine Mechanics and PhysicsChinese Academy of Sciences130033 ChangchunChina

The ability of an interferometer to characterize the spatial information of a light beam is often limited by the temporal profile of the beam,with femtosecond pulse characterization being particularly *** this study,we developed a simple,stable,controllable shearing and vectorial phase-shifting wedged reversal shearing interferometer that is able to characterize all types of coherent and partially coherent light *** proposed interferometer consists of only a single beam splitter cube with one wedged entrance face and is insensitive to environmental vibration due to its common path configuration.A near zero-path length difference of the proposed interferometer ensures its operation for ultrashort pulses,providing,for the first time,a simple and stable interferometric tool to fully characterize sub-100 fs laser *** common beam characterization can be carried out with the interferometer,such as the amplitude,phase,polarization,wavelength,and pulse ***,this technique is sensitive to the wavefront tilt and can be used for precise beam ***,this interferometer can be an essential tool for beam characterization,optical imaging,and the testing required for a wide range of applications,including astronomy,biomedicine,ophthalmology,optical testing and imaging systems,and adaptive optics.

关键词： interferometer ultrashort shearing

Theoretical study of substituent effect on the charge mobility of 2,5-bis(trialkylsilylethynyl)-1,1,3,4-tetraphenylsiloles

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China Chemistry 2010年第11期53卷 2311-2317页

作者： lam Jacky wing yip Department of Chemistry The Hong Kong University of Science & Technology

The theoretical calculation of the charge mobility of 2,5-bis(trialkylsilylethynyl)-1,1,3,4-tetraphenylsiloles is presented.B3LyP/6-31* calculations demonstrated that these silole molecules possessed large coupling matrix elements and reorganization energies for electron and hole transfers and high electron *** bulkiness of the trialkyl substituents influenced the charge mobility of the silole molecules,with the smaller trimethyl group imparting higher charge mobility than triethyl and triisopropyl substituents.

关键词： silole theoretical calculation substitution effect electronic structure charge mobility

Environmental threats to the Three Gorges Reservoir Region:Are mutagenic and genotoxic substances important?

在线全文

学校读者我要写书评

暂无评论

Journal of Environmental Sciences 2015年第12期27卷 172-174页

作者： Paul K.S.lam State Key Laboratory in Marine Pollution Department of Biology and ChemistryCity University of Hong Kong

The Three Gorges Dam project（TGDP）,with a total static investment of 95.46 billion RMB（US$10.97 billion）based on the 1993 price level,commenced in 1994 and was completed in *** creation of the Three Gorges Reservoir following the completion of the TGDP had brought about significant changes to the Three Gorges Reservoir Region（TGRR）,stretching from the town of Sandouping in Hubei Province to the Jiangjin District of Chongqing *** TGDP has led to progressive urbanization and industrialization of the TGRR,accompanied by increased shipping activities,greater

关键词： Three Gorges Dam Persistent organic pollutants yangtze River Sediment water

Effect of Eu^(3+),Bi^(3+),and Li^(+) doping on luminescent property of GdNbO_(4)

在线全文

学校读者我要写书评

暂无评论

Journal of Rare Earths 2024年第1期42卷 66-75,I0003页

作者： S.w.wi y.S.Lee Department of Physics and Integrative Institute of Basic Sciences Soongsil UniversitySeoul06978Republic of Korea

we investigated the emission properties of Eu^(3+)-doped GdNbO_(4)-based oxide phosphors,aiming at improving the Eu^(3+)emission by codoping of Bi^(3+)and Li^(+).Eu^(3+)singly doped,Eu^(3+)and Bi^(3+)doubly doped,and Eu^(3+),Bi^(3+),and Li^(+)triply doped GdNbO_(4)were prepared using a high-temperature solid-state reaction *** red-orange emissions from Eu^(3+)in the singly doped samples are significantly improved by the codoping of Bi^(3+),mainly due to the energy transfer from Bi^(3+)to Eu^(3+).The additional codoping of Li^(+)is found to increase the Eu^(3+)emission *** improvement might be attributed to the increase in the photoluminescent quantum yield originating from larger grain sizes and better crystallinity.A detailed analysis of the X-ray diffraction pattern and the asymmetric ratio estimated from the photoluminescence spectra show that the local lattice environment around Eu^(3+)in GdNbO_(4)does not change significantly with the codoping of Bi^(3+)and Li^(+).Our results indicate that Li^(+)doping is a promising way to improve the emission properties of rare-earth ion-doped GdNbO_(4)phosphors.

关键词： Eu^(3+) GdNbO_(4) Codoping of Li^(+) Photoluminescence Sensitizer Rare earths

On the Variability and Correlation of Surface Ozone and Carbon Monoxide Observed in Hong Kong Using Trajectory and Regression Analyses

在线全文

学校读者我要写书评

暂无评论

Advances in Atmospheric Sciences 2004年第1期21卷 141-152页

作者：王体健 K.S.lam C.w.TSANG S.C.KOT Department of Atmospheric Sciences Nanjing University Nanjing 210093 Environmental Engineering Unit Department of Civil & Structural Engineering The Hong Kong Polytechnic University Hong Kong Environmental Engineering Unit Department of Civil & Structural Engineering The Hong Kong Polytechnic University Hong Kong Department of Mechanical Engineering Hong Kong University of Science & Technology Hong Kong

This paper investigates, the variability and correlation of surface ozone (Os) and carbon monoxide (CO) observed at Cape D'Aguilar in Hong Kong from 1 January 1994 to 31 December 1995. Statistical analysis shows that the average O3 and CO mixing ratios during the two years are 32±17ppbv and 305±191 ppbv, respectively. The O3/CO ratio ranges from 0.05 to 0.6 ppbv/ppbv with its frequency peaking at 0.15. The raw dataset is divided into six groups using backward trajectory and cluster analyses. For data assigned to the same trajectory type, three groups are further sorted out based on CO and NOX mixing ratios. The correlation coefficients and slopes of O3/CO for the 18 groups are calculated using linear regression analysis. Finally, five kinds of air masses with different chemical features are identified: continental background (CB), marine background (MB), regional polluted continental (RPC), perturbed marine (P*M), and local polluted (LP) air masses. Further studies indicate that O3 and CO in the continental and marine background air masses (CB and MB) are positively correlated for the reason that they are well mixed over the long range transport before arriving at the site. The negative correlation between O3 and CO in air mass LP is believed to be associated with heavy anthropogenic influence, which results from the enhancement by local sources as indicated by high CO and NOx and depletion of O3 when mixed with fresh emissions. The positive correlation in the perturbed marine air mass P*M favors the low photochemical production of O3. The negative correlation found in the regional polluted continental air mass RPC is different from the observations at Oki Island in Japan due to the more complex O3 chemistry at Cape D'Aguilar.

关键词： ozone carbon monoxide trajectory

Ecoregional height-diameter models for Scots pine in Turkiye

在线全文

学校读者我要写书评

暂无评论

Journal of Forestry Research 2024年第6期35卷 49-61页

作者： Fadime Sağlam Oytun Emre Sakici Department of Forest Engineering Faculty of ForestryKastamonu University37150 KastamonuTurkey

Ecoregion-based height-diameter models were developed in the present study for Scots pine(Pinus sylves-tris L.)stands in Turkiye and included several ecological factors derived from a pre-existing ecoregional classification *** data were obtained from 2831 sample trees in 292 sample *** generalized height–diameter models were developed,and the best model(HD10)was selected according to statistical ***,nonlinear mixed-effects modeling was applied to the best *** R2 for the generalized height‒diameter model(Richards function)modified by Sharma and Parton is 0.951,and the final model included number of trees,dominant height,and diameter at breast height,with a random parameter associated with each ecoregion attached to the inverse of the mean basal *** full model predictions using the nonlinear mixed-effects model and the reduced model(HD10)predictions were compared using the nonlinear sum of extra squares test,which revealed significant differences between ecore-gions;ecoregion-based height–diameter models were thus found to be suitable to *** addition,using these models in appropriate ecoregions was very important for achieving reliable predictions with low prediction errors.

关键词： Tree height Nonlinear mixed-effects modelling Nonlinear sum of extra squares method Ecoregion Scots pine