A5103094738- Statistical Methods and Inference
- Sparse and Compressive Sensing Techniques
- Advanced Graph Neural Networks
- Advanced Optimization Algorithms Research
- Climate change and permafrost
- Cryospheric studies and observations
- Nuclear materials and radiation effects
- Climate variability and models
- Bayesian Methods and Mixture Models
- Atmospheric chemistry and aerosols
- Arctic and Antarctic ice dynamics
- Optimization and Variational Analysis
- Bone and Joint Diseases
- Risk and Portfolio Optimization
- Computational Fluid Dynamics and Aerodynamics
- Gene expression and cancer classification
- Spectroscopy and Chemometric Analyses
- Graph Theory and Algorithms
- Numerical methods for differential equations
- Systemic Lupus Erythematosus Research
- Advanced Clustering Algorithms Research
- Atmospheric Ozone and Climate
- Statistical Methods and Bayesian Inference
- Remote Sensing in Agriculture
- China's Socioeconomic Reforms and Governance
Dawu County People's Hospital
2025
Chinese Academy of Sciences
2009-2024
Academy of Mathematics and Systems Science
2020-2024
Institute of Applied Mathematics
2024
National University of Singapore
2018-2021
Hong Kong Polytechnic University
2020-2021
Dongguan University of Technology
2020-2021
Changchun Institute of Applied Chemistry
2009-2019
Institute of Tibetan Plateau Research
2009-2019
University of Science and Technology of China
2019
Abstract. Two ice cores were retrieved from high elevations (~5800 m a.s.l.) at Mt. Nyainqêntanglha and Geladaindong in the southern central Tibetan Plateau region. The combined tracer analysis of tritium (3H), 210Pb mercury, along with other chemical records, provided multiple lines evidence supporting that two coring sites had not received net accumulation since least 1950s 1980s, respectively. These results implied an annual loss rate more than several hundred millimeter water equivalent...
Abstract Understanding past atmospheric dust variability is necessary to put modern into historical context and assess the impacts of on climate. In Asia, meteorological data temporally limited, beginning only in 1950s. High‐resolution ice cores provide ideal archive for reconstructing preinstrumental concentrations. Using a ~500 year (1477–1982 A.D.) annually resolved calcium (Ca) proxy from Tibetan Plateau (TP) core, we demonstrate lowest concentrations years during latter twentieth...
Terrestrial net primary production (NPP), the balance of gross (GPP) and autotrophic respiration (AR), is a critical measure carbon sequestration capacity for Earth's land surface. The aim this study was to understand spatio-temporal variability NPP associated with GPP AR in Yangtze River Basin (YRB), China, from 2000 2009 during which basin warmed significantly. We first derived carbon-use efficiency (CUE) improved Moderate Resolution Imaging Spectroradiometer GPP/NPP products (MOD17) then...
In autumn 2005, a joint expedition between the University of Maine and Institute Tibetan Plateau Research recovered three ice cores from Guoqu Glacier (33°34′37.8″N, 91°10′35.3″E, 5720 m above sea level) on northern side Mt. Geladaindong, central Plateau. Isotopes ( δ 18 O), major soluble ions (Na + , K Mg 2+ Ca Cl − NO 3 SO 4 2− ), radionuclide β ‐activity) measurements one revealed 70‐year record (1935–2005). Statistical analysis ion time series suggests that atmospheric dust species...
Abstract Plasma‐sprayed 8YSZ (zirconia stabilized with 8 wt% yttria)/NiCoCrAlYTa thermal barrier coatings (TBCs) were laser‐glazed using a continuous‐wave CO 2 laser. Open pores within the coating surface eliminated and an external densified layer was generated by laser‐glazing. The hot corrosion resistances of plasma‐sprayed investigated. two specimens exposed for same period 100 h at 900 °C to salt mixture vanadium pentoxide (V O 5 ) sodium sulfate (Na SO 4 ). Serious crack spallation...
.Strong variational sufficiency is a newly proposed property, which turns out to be of great use in the convergence analysis multiplier methods. However, what this property implies for nonpolyhedral problems remains puzzle. In paper, we prove equivalence between strong and second-order sufficient condition (SOSC) nonlinear semidefinite programming (NLSDP) without requiring uniqueness or any other constraint qualifications. Based on characterization, local augmented Lagrangian method (ALM)...
Undirected graphical models have been especially popular for learning the conditional independence structure among a large number of variables where observations are drawn independently and identically from same distribution. However, many modern statistical problems would involve categorical data or time-varying data, which might follow different but related underlying distributions. In order to learn collection simultaneously, various joint inducing sparsity in graphs similarity across...
We consider the problem of learning a graph under Laplacian constraint with non-convex penalty: minimax concave penalty (MCP). For solving MCP penalized graphical model, we design an inexact proximal difference-of-convex algorithm (DCA) and prove its convergence to critical points. note that each subproblem DCA enjoys nice property objective function in dual is continuously differentiable semismooth gradient. Therefore, apply efficient Newton method subproblems DCA. Numerical experiments on...
We consider the problem of jointly learning row-wise and column-wise dependencies matrix-variate observations, which are modelled separately by two precision matrices. Due to complicated structure Kronecker-product matrices in commonly used Gaussian graphical models, a sparser Kronecker-sum was proposed recently based on Cartesian product graphs. However, existing methods for estimating structured do not scale well large datasets. In this paper, we introduce DNNLasso, diagonally non-negative...
Common clustering methods, such as $k$-means and convex clustering, group similar vector-valued observations into clusters. However, with the increasing prevalence of matrix-valued observations, which often exhibit low rank characteristics, there is a growing need for specialized techniques these data types. In this paper, we propose model tailored observations. Our approach extends originally designed to classify Additionally, it serves relaxation method proposed by Z. Lyu, D. Xia...
Recently, the square root principal component pursuit (SRPCP) model has garnered significant research interest. It is shown in literature that SRPCP guarantees robust matrix recovery with a universal, constant penalty parameter. While its statistical advantages are well-documented, computational aspects from an optimization perspective remain largely unexplored. In this paper, we focus on developing efficient algorithms for solving problem. Specifically, propose tuning-free alternating...
This article analyses the non-market value of cultivated land resources firstly, then makes a further study why thecultivated lack compensation under current expropriation systems. Theresults prove that vacancy not only does harm to rights farmers, but also lossof total social welfare and results in disorder circulation during non-agricultural. Finally, this paper tries toimprove perfect systems based on category.
We consider the problem of learning a graph under Laplacian constraint with non-convex penalty: minimax concave penalty (MCP). For solving MCP penalized graphical model, we design an inexact proximal difference-of-convex algorithm (DCA) and prove its convergence to critical points. note that each subproblem DCA enjoys nice property objective function in dual is continuously differentiable semismooth gradient. Therefore, apply efficient Newton method subproblems DCA. Numerical experiments on...
In this paper we study the computation of nonparametric maximum likelihood estimator (NPMLE) in multivariate mixture models. Our first approach discretizes infinite dimensional convex optimization problem by fixing support points NPMLE and optimizing over proportions. context propose, leveraging sparsity solution, an efficient scalable semismooth Newton based augmented Lagrangian method (ALM). algorithm beats state-of-the-art methods~\cite{koenker2017rebayes, kim2020fast} can handle $n...
An Efficient Linearly Convergent Regularized Proximal Point Algorithm for Fused Multiple Graphical Lasso Problems
Abstract. The Upper Indus River Basin (UIB) has developed the largest midlatitude mountain glaciers worldwide. Ice thickness and volume distribution are important prerequisites for glaciological hydrological investigations. In this paper, we presented detailed estimates of ice in UIB region. Using ground penetrating radar, measured glacier on six typical glaciers; obtained parameters GlabTOP2 from these measurements analyzed its uncertainty. verified model, simulated subcatchments. results...
Square-root (loss) regularized models have recently become popular in linear regression due to their nice statistical properties. Moreover, some of these can be interpreted as the distributionally robust optimization counterparts traditional least-squares models. In this paper, we give a unified proof show that any square-root model whose penalty function being sum simple norm and seminorm (DRO) formulation corresponding problem. particular, optimal transport cost DRO is given by certain...
The Riemannian Augmented Lagrangian Method (RALM), a recently proposed algorithm for nonsmooth optimization problems on manifolds, has consistently exhibited high efficiency as evidenced in prior studies \cite{ZBDZ21,ZBD22}. It often demonstrates rapid local linear convergence rate. However, comprehensive analysis of the RALM under more realistic assumptions, notably without imposition uniqueness assumption multiplier, remains an uncharted territory. In this paper, we introduce manifold...