oalib

Publish in OALib Journal

ISSN: 2333-9721

APC: Only $99

Submit

Any time

2020 ( 2 )

2019 ( 304 )

2018 ( 1901 )

2017 ( 1799 )

Custom range...

Search Results: 1 - 10 of 120068 matches for " Jialei Wang "
All listed articles are free for downloading (OA Articles)
Page 1 /120068
Display every page Item
Inference for Sparse Conditional Precision Matrices
Jialei Wang,Mladen Kolar
Statistics , 2014,
Abstract: Given $n$ i.i.d. observations of a random vector $(X,Z)$, where $X$ is a high-dimensional vector and $Z$ is a low-dimensional index variable, we study the problem of estimating the conditional inverse covariance matrix $\Omega(z) = (E[(X-E[X \mid Z])(X-E[X \mid Z])^T \mid Z=z])^{-1}$ under the assumption that the set of non-zero elements is small and does not depend on the index variable. We develop a novel procedure that combines the ideas of the local constant smoothing and the group Lasso for estimating the conditional inverse covariance matrix. A proximal iterative smoothing algorithm is used to solve the corresponding convex optimization problems. We prove that our procedure recovers the conditional independence assumptions of the distribution $X \mid Z$ with high probability. This result is established by developing a uniform deviation bound for the high-dimensional conditional covariance matrix from its population counterpart, which may be of independent interest. Furthermore, we develop point-wise confidence intervals for individual elements of the conditional inverse covariance matrix. We perform extensive simulation studies, in which we demonstrate that our proposed procedure outperforms sensible competitors. We illustrate our proposal on a S&P 500 stock price data set.
Bayesian optimization for materials design
Peter I. Frazier,Jialei Wang
Mathematics , 2015,
Abstract: We introduce Bayesian optimization, a technique developed for optimizing time-consuming engineering simulations and for fitting machine learning models on large datasets. Bayesian optimization guides the choice of experiments during materials design and discovery to find good material designs in as few experiments as possible. We focus on the case when materials designs are parameterized by a low-dimensional vector. Bayesian optimization is built on a statistical technique called Gaussian process regression, which allows predicting the performance of a new design based on previously tested designs. After providing a detailed introduction to Gaussian process regression, we introduce two Bayesian optimization methods: expected improvement, for design problems with noise-free evaluations; and the knowledge-gradient method, which generalizes expected improvement and may be used in design problems with noisy evaluations. Both methods are derived using a value-of-information analysis, and enjoy one-step Bayes-optimality.
Fire and Explosion Hazard Prediction Base on Virtual Reality in Tank Farm
Jialei Tan,Yushu Xie,Tong Wang
Journal of Software , 2012, DOI: 10.4304/jsw.7.3.678-682
Abstract: Many fires and explosions happen in tank farm. A fire and explosion simulation method based on VR is used. Tank fire and explosion special effect is realized by classical particle system. The real-time fire growth model, fireball growth model, fireball heat radiation model are created. The shock wave model of BLEVE also is created. The models and results of FDS (Fire Dynamic Simulator) are built in virtual reality system. It can show the process of fire and explosion development and calculate the hazard of fire and explosion. The virtual reality system can predict and provide a method to control the fire and explosion hazard.
Distributed Multitask Learning
Jialei Wang,Mladen Kolar,Nathan Srebro
Computer Science , 2015,
Abstract: We consider the problem of distributed multi-task learning, where each machine learns a separate, but related, task. Specifically, each machine learns a linear predictor in high-dimensional space,where all tasks share the same small support. We present a communication-efficient estimator based on the debiased lasso and show that it is comparable with the optimal centralized method.
Effect of Cultivation Pattern on the Light Radiation of Group Canopy and Yield of Spring Soybean (Glycine Max L. Merrill)  [PDF]
Jialei Xiao, uijiang Wang, Ming Zhao, Jing Yin, Wei Li, Wan Li, Yongcai Lai, Xiatian Shu, Yang Zhao, Yingdong Bi
American Journal of Plant Sciences (AJPS) , 2013, DOI: 10.4236/ajps.2013.46148
Abstract:


Heilongjiang Province is the main soybean-producing area in china. In this study, we analyzed the canopy structure, dynamic characteristics of light radiation and yield of Hefeng50 (the main variety of soybean in Heilongjiang Province) under six different cultivation patterns (ORP, TPCR, ORCP, BRHD, SRHD and FPHD). The results showed that SRHD and BRHD at different growth period (blossom period R1, podding R3 and grain filing period R5) produced an even distribution of the population leaf area, suitable mean foliage inclination angle (MFIA), low transparency coefficients for defuse penetration (TCDP) and transparency coefficients for radiation penetration (TCRP), high leaf area index (LAI), extinction light coefficient (K value), fraction of radiation intercepted (FRI) and light energy utilization rate. Grain number, dry matter weight per plant, and yield of SRHD and BRHD were significantly higher than those of other cultivation patterns. The yield of SRHD, BRHD, ORCP, FPHD and TPCR was increased by 136%, 112%, 79%, 50.1% and 14.7%, respectively, compared to that of ORP. These results suggest that SRHD and BRHD are the optimal cultivation pattern for the improvement of soybean yield in phaeozem region of northeastern China.


Exact Soft Confidence-Weighted Learning
Jialei Wang,Peilin Zhao,Steven C. H. Hoi
Computer Science , 2012,
Abstract: In this paper, we propose a new Soft Confidence-Weighted (SCW) online learning scheme, which enables the conventional confidence-weighted learning method to handle non-separable cases. Unlike the previous confidence-weighted learning algorithms, the proposed soft confidence-weighted learning method enjoys all the four salient properties: (i) large margin training, (ii) confidence weighting, (iii) capability to handle non-separable data, and (iv) adaptive margin. Our experimental results show that the proposed SCW algorithms significantly outperform the original CW algorithm. When comparing with a variety of state-of-the-art algorithms (including AROW, NAROW and NHERD), we found that SCW generally achieves better or at least comparable predictive accuracy, but enjoys significant advantage of computational efficiency (i.e., smaller number of updates and lower time cost).
Fabrication of three-dimensional microdisk resonators in calcium fluoride by femtosecond laser micromachining
Jintian Lin,Yingxin Xu,Jialei Tang,Nengwen Wang,Jiangxin Song,Fei He,Wei Fang,Ya Cheng
Physics , 2014, DOI: 10.1007/s00339-014-8388-1
Abstract: We report on fabrication of on-chip calcium fluoride (CaF2) microdisk resonators using water-assisted femtosecond laser micromachining. Focused ion beam (FIB) milling is used to create ultra-smooth sidewalls. The quality (Q)-factors of the fabricated microresonators are measured to be 4.2x10^4 at wavelengths near 1550 nm. The Q factor is mainly limited by the scattering from the bottom surface of the disk whose roughness remains high due to the femtosecond laser micromachining process. This technique facilitates formation of on-chip microresonators on various kinds of bulk crystalline materials, which can benefit a wide range of applications such as nonlinear optics, quantum optics, and chip-level integration of photonic devices.
On-chip tuning of the resonant wavelength in a high-Q microresonator integrated with a microheater
Jialei Tang,Jintian Lin,Jiangxin Song,Zhiwei Fang,Min Wang,Yang Liao,Lingling Qiao,Ya Cheng
Physics , 2014,
Abstract: We report on fabrication of a microtoroid resonator of high-quality (high-Q) factor integrated with an on-chip microheater. Both the microresonator and microheater are fabricated using femtosecond laser three-dimensional (3D) micromachining. The microheater, which is located about 200 micron away from the microresonator, has a footprint size of 200 micron by 400 micron. Tuning of the resonant wavelength in the microresonator has been achieved by varying the voltage applied on the microheater. The drifting of the resonant wavelength shows a linear dependence on the square of the voltage applied on the microheater. We found that the response time of the microresonator is less than 10 secs which is significantly shorter than the time required for reaching a thermal equilibrium on conventional heating instruments such as an external electric heater.
Fast Bounded Online Gradient Descent Algorithms for Scalable Kernel-Based Online Learning
Peilin Zhao,Jialei Wang,Pengcheng Wu,Rong Jin,Steven C. H. Hoi
Computer Science , 2012,
Abstract: Kernel-based online learning has often shown state-of-the-art performance for many online learning tasks. It, however, suffers from a major shortcoming, that is, the unbounded number of support vectors, making it non-scalable and unsuitable for applications with large-scale datasets. In this work, we study the problem of bounded kernel-based online learning that aims to constrain the number of support vectors by a predefined budget. Although several algorithms have been proposed in literature, they are neither computationally efficient due to their intensive budget maintenance strategy nor effective due to the use of simple Perceptron algorithm. To overcome these limitations, we propose a framework for bounded kernel-based online learning based on an online gradient descent approach. We propose two efficient algorithms of bounded online gradient descent (BOGD) for scalable kernel-based online learning: (i) BOGD by maintaining support vectors using uniform sampling, and (ii) BOGD++ by maintaining support vectors using non-uniform sampling. We present theoretical analysis of regret bound for both algorithms, and found promising empirical performance in terms of both efficacy and efficiency by comparing them to several well-known algorithms for bounded kernel-based online learning on large-scale datasets.
Molecular Subtype Classification Is a Determinant of Non-Sentinel Lymph Node Metastasis in Breast Cancer Patients with Positive Sentinel Lymph Nodes
Wenbin Zhou, Zhongyuan He, Jialei Xue, Minghai Wang, Xiaoming Zha, Lijun Ling, Lin Chen, Shui Wang, Xiaoan Liu
PLOS ONE , 2012, DOI: 10.1371/journal.pone.0035881
Abstract: Background Previous studies suggested that the molecular subtypes were strongly associated with sentinel lymph node (SLN) status. The purpose of this study was to determine whether molecular subtype classification was associated with non-sentinel lymph nodes (NSLN) metastasis in patients with a positive SLN. Methodology and Principal Findings Between January 2001 and March 2011, a total of 130 patients with a positive SLN were recruited. All these patients underwent a complete axillary lymph node dissection. The univariate and multivariate analyses of NSLN metastasis were performed. In univariate and multivariate analyses, large tumor size, macrometastasis and high tumor grade were all significant risk factors of NSLN metastasis in patients with a positive SLN. In univariate analysis, luminal B subgroup showed higher rate of NSLN metastasis than other subgroup (P = 0.027). When other variables were adjusted in multivariate analysis, the molecular subtype classification was a determinant of NSLN metastasis. Relative to triple negative subgroup, both luminal A (P = 0.047) and luminal B (P = 0.010) subgroups showed a higher risk of NSLN metastasis. Otherwise, HER2 over-expression subgroup did not have a higher risk than triple negative subgroup (P = 0.183). The area under the curve (AUC) value was 0.8095 for the Cambridge model. When molecular subtype classification was added to the Cambridge model, the AUC value was 0.8475. Conclusions Except for other factors, molecular subtype classification was a determinant of NSLN metastasis in patients with a positive SLN. The predictive accuracy of mathematical models including molecular subtype should be determined in the future.
Page 1 /120068
Display every page Item


Home
Copyright © 2008-2017 Open Access Library. All rights reserved.