In this work,we have proposed a generative model,called VAE-KRnet,for density estimation or approximation,which combines the canonical variational autoencoder(VAE)with our recently developed flow-based generativemodel...In this work,we have proposed a generative model,called VAE-KRnet,for density estimation or approximation,which combines the canonical variational autoencoder(VAE)with our recently developed flow-based generativemodel,called KRnet.VAE is used as a dimension reduction technique to capture the latent space,and KRnet is used to model the distribution of the latent variable.Using a linear model between the data and the latent variable,we show that VAE-KRnet can be more effective and robust than the canonical VAE.VAE-KRnet can be used as a density model to approximate either data distribution or an arbitrary probability density function(PDF)known up to a constant.VAE-KRnet is flexible in terms of dimensionality.When the number of dimensions is relatively small,KRnet can effectively approximate the distribution in terms of the original random variable.For high-dimensional cases,we may use VAE-KRnet to incorporate dimension reduction.One important application of VAE-KRnet is the variational Bayes for the approximation of the posterior distribution.The variational Bayes approaches are usually based on the minimization of the Kullback-Leibler(KL)divergence between the model and the posterior.For highdimensional distributions,it is very challenging to construct an accurate densitymodel due to the curse of dimensionality,where extra assumptions are often introduced for efficiency.For instance,the classical mean-field approach assumes mutual independence between dimensions,which often yields an underestimated variance due to oversimplification.To alleviate this issue,we include into the loss the maximization of the mutual information between the latent random variable and the original random variable,which helps keep more information from the region of low density such that the estimation of variance is improved.Numerical experiments have been presented to demonstrate the effectiveness of our model.展开更多
Interest in automated data classification and identification systems has increased over the past years in conjunction with the high demand for artificial intelligence and security applications.In particular,recognizin...Interest in automated data classification and identification systems has increased over the past years in conjunction with the high demand for artificial intelligence and security applications.In particular,recognizing human activities with accurate results have become a topic of high interest.Although the current tools have reached remarkable successes,it is still a challenging problem due to various uncontrolled environments and conditions.In this paper two statistical frameworks based on nonparametric hierarchical Bayesian models and Gamma distribution are proposed to solve some realworld applications.In particular,two nonparametric hierarchical Bayesian models based on Dirichlet process and Pitman-Yor process are developed.These models are then applied to address the problem of modelling grouped data where observations are organized into groups and these groups are statistically linked by sharing mixture components.The choice of the Gamma mixtures is motivated by its flexibility for modelling heavy-tailed distributions.In addition,deploying the Dirichlet process prior is justified by its advantage of automatically finding the right number of components and providing nice properties.Moreover,a learning step via variational Bayesian setting is presented in a flexible way.The priors over the parameters are selected appropriately and the posteriors are approximated effectively in a closed form.Experimental results based on a real-life applications that concerns texture classification and human actions recognition show the capabilities and effectiveness of the proposed framework.展开更多
The seasonal variations in phytoplankton community structure were investigated for the Sanggou Bay （SGB） and the adjacent Ailian Bay （ALB） and Lidao Bay （LDB） in Shandong Peninsula,eastern China.The species comp...The seasonal variations in phytoplankton community structure were investigated for the Sanggou Bay （SGB） and the adjacent Ailian Bay （ALB） and Lidao Bay （LDB） in Shandong Peninsula,eastern China.The species composition and cell abundance of phytoplankton in the bay waters in spring （April 2011）,summer （August 2011）,autumn （October 2011）,and winter （January 2012） were examined using the Uterm6hl method.A total of 80 taxa of phytoplankton that belong to 39 genera of 3 phyla were identified.These included 64 species of 30 genera in the Phylum Bacillariophyta,13 species of 8 genera in the Phylum Dinophyta,and 3 species of 1 genus in the Phylum Chrysophyta.During the four seasons,the number of phytoplankton species （43） was the highest in spring,followed by summer and autumn （40）,and the lowest number ofphytoplankton species （35） was found in winter.Diatoms,especially Paralia sulcata （Ehrenberg） Cleve and Coscinodiscus oculus-iridis Ehrenberg,were predominant in the phytoplankton community throughout the study period,whereas the dominance of dinoflagellate appeared in summer only.The maximum cell abundance of phytoplankton was detected in summer （average 8.08 × 103 cells L-1） whereas their minimum abundance was found in autumn （average 2.60 x 103 cellsL-1）.The phytoplankton abundance was generally higher in the outer bay than in the inner bay in spring and autumn.In summer,the phytoplankton cells were mainly concentrated in the south of inner SGB,with peak abundance observed along the western coast.In winter,the distribution of phytoplankton cells showed 3 patches,with peak abundance along the western coast as well.On seasonal average,the Shannon-Wiener diversity indices of phytoplankton community ranged from 1.17 to 1.78 （autumn 〉 summer 〉 spring 〉 winter）,and the Pielou＇s evenness indices of phytoplankton ranged from 0.45 to 0.65 （autumn 〉 spring 〉 summer〉 winter）.According to the results of canonical correspondence analysis,phosphate level was the major factor that limited the occurrence of P.sulcata and C.oculus-iridis,whereas optimal temperature and low salinity were responsible for Prorocentrum blooms in summer.The detailed description of seasonal variations in phytoplankton community structure in the three bays provide reference data for future studies on marine ecosystems and mariculture in adjacent areas.展开更多
Based on investigation data of PHC content in Jiaozhou Bay,China from 1979 to 1983,the seasonal variations of PHC content and monthly changes of precipitation in Jiaozhou Bay were analyzed. The results showed that see...Based on investigation data of PHC content in Jiaozhou Bay,China from 1979 to 1983,the seasonal variations of PHC content and monthly changes of precipitation in Jiaozhou Bay were analyzed. The results showed that seen from the spatial and temporal distribution,the seasonal variation of PHC content in the surface water of Jiaozhou Bay was based on the flow of the rivers as well as human activity,so PHC content in the rivers depended on the flow of the rivers and human activity,and the peaks and valleys of PHC content appeared in various seasons. The seasonal variation of PHC content in the surface water of Jiaozhou Bay depended on its land transfer process. The land transfer process was composed of use of PHC by mankind,deposition of PHC in soil and on the earth＇s surface,and transportation of PHC to offshore waters of sea by rivers and surface runoff. PHC content depended on mankind during the process from being used to entering soil and on precipitation during the process of being transported from soil to ocean.展开更多
A novel scheme to joint phase noise （PHN） correcting and channel noise variance estimating for orthogonal frequency division multiplexing （OFDM） signal was proposed, The new scheme was based on the variational Bay...A novel scheme to joint phase noise （PHN） correcting and channel noise variance estimating for orthogonal frequency division multiplexing （OFDM） signal was proposed, The new scheme was based on the variational Bayes （VB） method and discrete cosine transform （DCT） approximation. Compared with the least squares （LS） based scheme, the proposed scheme could overcome the over-fitting phenomenon and thus lead to an improved performance. Computer simulations showed that the proposed VB based scheme outperforms the existing LS based scheme展开更多
基金X.Wan has been supported by NSF grant DMS-1913163S.Wei has been supported by NSF grant ECCS-1642991.
文摘In this work,we have proposed a generative model,called VAE-KRnet,for density estimation or approximation,which combines the canonical variational autoencoder(VAE)with our recently developed flow-based generativemodel,called KRnet.VAE is used as a dimension reduction technique to capture the latent space,and KRnet is used to model the distribution of the latent variable.Using a linear model between the data and the latent variable,we show that VAE-KRnet can be more effective and robust than the canonical VAE.VAE-KRnet can be used as a density model to approximate either data distribution or an arbitrary probability density function(PDF)known up to a constant.VAE-KRnet is flexible in terms of dimensionality.When the number of dimensions is relatively small,KRnet can effectively approximate the distribution in terms of the original random variable.For high-dimensional cases,we may use VAE-KRnet to incorporate dimension reduction.One important application of VAE-KRnet is the variational Bayes for the approximation of the posterior distribution.The variational Bayes approaches are usually based on the minimization of the Kullback-Leibler(KL)divergence between the model and the posterior.For highdimensional distributions,it is very challenging to construct an accurate densitymodel due to the curse of dimensionality,where extra assumptions are often introduced for efficiency.For instance,the classical mean-field approach assumes mutual independence between dimensions,which often yields an underestimated variance due to oversimplification.To alleviate this issue,we include into the loss the maximization of the mutual information between the latent random variable and the original random variable,which helps keep more information from the region of low density such that the estimation of variance is improved.Numerical experiments have been presented to demonstrate the effectiveness of our model.
基金The authors would like to thank Taif University Researchers Supporting Project number(TURSP-2020/26),Taif University,Taif,Saudi ArabiaThey would like also to thank Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2022R40),Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Interest in automated data classification and identification systems has increased over the past years in conjunction with the high demand for artificial intelligence and security applications.In particular,recognizing human activities with accurate results have become a topic of high interest.Although the current tools have reached remarkable successes,it is still a challenging problem due to various uncontrolled environments and conditions.In this paper two statistical frameworks based on nonparametric hierarchical Bayesian models and Gamma distribution are proposed to solve some realworld applications.In particular,two nonparametric hierarchical Bayesian models based on Dirichlet process and Pitman-Yor process are developed.These models are then applied to address the problem of modelling grouped data where observations are organized into groups and these groups are statistically linked by sharing mixture components.The choice of the Gamma mixtures is motivated by its flexibility for modelling heavy-tailed distributions.In addition,deploying the Dirichlet process prior is justified by its advantage of automatically finding the right number of components and providing nice properties.Moreover,a learning step via variational Bayesian setting is presented in a flexible way.The priors over the parameters are selected appropriately and the posteriors are approximated effectively in a closed form.Experimental results based on a real-life applications that concerns texture classification and human actions recognition show the capabilities and effectiveness of the proposed framework.
基金supported by the National Program on Key Basic Research Project of China (Grant Nos. 2011CB409804 and 2015CB954002)Program for New Century Excellent Talents in University (NCET-12-1065)the National Natural Science Foundation of China (Grant No.41176136) to J. Sun
文摘The seasonal variations in phytoplankton community structure were investigated for the Sanggou Bay （SGB） and the adjacent Ailian Bay （ALB） and Lidao Bay （LDB） in Shandong Peninsula,eastern China.The species composition and cell abundance of phytoplankton in the bay waters in spring （April 2011）,summer （August 2011）,autumn （October 2011）,and winter （January 2012） were examined using the Uterm6hl method.A total of 80 taxa of phytoplankton that belong to 39 genera of 3 phyla were identified.These included 64 species of 30 genera in the Phylum Bacillariophyta,13 species of 8 genera in the Phylum Dinophyta,and 3 species of 1 genus in the Phylum Chrysophyta.During the four seasons,the number of phytoplankton species （43） was the highest in spring,followed by summer and autumn （40）,and the lowest number ofphytoplankton species （35） was found in winter.Diatoms,especially Paralia sulcata （Ehrenberg） Cleve and Coscinodiscus oculus-iridis Ehrenberg,were predominant in the phytoplankton community throughout the study period,whereas the dominance of dinoflagellate appeared in summer only.The maximum cell abundance of phytoplankton was detected in summer （average 8.08 × 103 cells L-1） whereas their minimum abundance was found in autumn （average 2.60 x 103 cellsL-1）.The phytoplankton abundance was generally higher in the outer bay than in the inner bay in spring and autumn.In summer,the phytoplankton cells were mainly concentrated in the south of inner SGB,with peak abundance observed along the western coast.In winter,the distribution of phytoplankton cells showed 3 patches,with peak abundance along the western coast as well.On seasonal average,the Shannon-Wiener diversity indices of phytoplankton community ranged from 1.17 to 1.78 （autumn 〉 summer 〉 spring 〉 winter）,and the Pielou＇s evenness indices of phytoplankton ranged from 0.45 to 0.65 （autumn 〉 spring 〉 summer〉 winter）.According to the results of canonical correspondence analysis,phosphate level was the major factor that limited the occurrence of P.sulcata and C.oculus-iridis,whereas optimal temperature and low salinity were responsible for Prorocentrum blooms in summer.The detailed description of seasonal variations in phytoplankton community structure in the three bays provide reference data for future studies on marine ecosystems and mariculture in adjacent areas.
文摘Based on investigation data of PHC content in Jiaozhou Bay,China from 1979 to 1983,the seasonal variations of PHC content and monthly changes of precipitation in Jiaozhou Bay were analyzed. The results showed that seen from the spatial and temporal distribution,the seasonal variation of PHC content in the surface water of Jiaozhou Bay was based on the flow of the rivers as well as human activity,so PHC content in the rivers depended on the flow of the rivers and human activity,and the peaks and valleys of PHC content appeared in various seasons. The seasonal variation of PHC content in the surface water of Jiaozhou Bay depended on its land transfer process. The land transfer process was composed of use of PHC by mankind,deposition of PHC in soil and on the earth＇s surface,and transportation of PHC to offshore waters of sea by rivers and surface runoff. PHC content depended on mankind during the process from being used to entering soil and on precipitation during the process of being transported from soil to ocean.
基金supported by the National Basic Research Program of China (2009CB320401)the National Key Scientific and Technological Project of China (2012ZX03004005-002,2010ZX03003-001)the National Natural Science Foundation of China (61171099)
文摘A novel scheme to joint phase noise （PHN） correcting and channel noise variance estimating for orthogonal frequency division multiplexing （OFDM） signal was proposed, The new scheme was based on the variational Bayes （VB） method and discrete cosine transform （DCT） approximation. Compared with the least squares （LS） based scheme, the proposed scheme could overcome the over-fitting phenomenon and thus lead to an improved performance. Computer simulations showed that the proposed VB based scheme outperforms the existing LS based scheme
