In this paper, we propose a data-adaptive non-parametric kernel learning framework in margin based kernel methods. In model formulation, given an initial kernel matrix, a data-adaptive matrix with two constraints is i...
详细信息
In this paper, we propose a data-adaptive non-parametric kernel learning framework in margin based kernel methods. In model formulation, given an initial kernel matrix, a data-adaptive matrix with two constraints is imposed in an entry-wise scheme. Learning this data-adaptive matrix in a formulation-free strategy enlarges the margin between classes and thus improves the model flexibility. The introduced two constraints are imposed either exactly (on small data sets) or approximately (on large data sets) in our model, which provides a controllable trade-off between model flexibility and complexity with theoretical demonstration. In algorithm optimization, the objective function of our learning framework is proven to be gradient-Lipschitz continuous. Thereby, kernel and classifier/regressor learning can be efficiently optimized in a unified framework via Nesterov's acceleration. For the scalability issue, we study a decomposition-based approach to our model in the large sample case. The effectiveness of this approximation is illustrated by both empirical studies and theoretical guarantees. Experimental results on various classification and regression benchmark data sets demonstrate that our non-parametric kernel learning framework achieves good performance when compared with other representative kernel learning based algorithms.
Geo-spatial ontologies can provide a formal description of concepts, relationships, activities, features and rules in GIS domain. However, simply use them only allows to partially solve semantic conflicts, and does no...
详细信息
Image saliency detection is an active research topic in the community of computer vision and multimedia. Fusing complementary RGB and thermal infrared data has been proven to be effective for image saliency detection....
详细信息
The atlas of human acupoints and meridians has been utilized in clinical practice for almost a millennium although the anatomical structures and functions remain to be clarified. It has recently been reported that a l...
详细信息
The atlas of human acupoints and meridians has been utilized in clinical practice for almost a millennium although the anatomical structures and functions remain to be clarified. It has recently been reported that a long-distance interstitial fluid (ISF) circulatory pathway may originate from the acupoints in the extremities. As observed in living human subjects, cadavers and animals using magnetic resonance imaging and fluorescent tracers, the ISF flow pathways include at least 4 types of anatomical structures: the cutaneous-, perivenous-, periarterial-, and neural-pathways. Unlike the blood or lymphatic vessels, these ISF flow pathways are composed of highly ordered and topologically connected interstitial fibrous connective tissues that may work as guiderails for the ISF to flow actively over long distance under certain driving forces. Our experimental results demonstrated that most acupoints in the extremity endings connect with one or more ISF flow pathways and comprise a complex network of acupoint-ISF-pathways. We also found that this acupoint-ISF-pathway network can connect to visceral organs or tissues such as the pericardium and epicardium, even though the topographical geometry in human extremities does not totally match the meridian lines on the atlas that is currently used in traditional Chinese medicine. Based on our experimental data, the following working hypotheses are proposed: 1, there are one or more ISF flow pathways, including at least one cutaneous pathway, originated from an acupoint on the body surface. 2, the acupoints on the body surface specifically connect with certain visceral organs or tissues via ISF flow. And 3, the acupoint-originated ISF pathways constitute a complex connective network and can modulate the ISF and bio-signals in the microenvironments around cells in certain visceral organs or tissues from body surfaces. A comprehensive atlas will be constructed to systemically reveal the detailed anatomical structures of the acupoi
Dielectric ceramics with high energy storage density and energy efficiency play an important role in high power energy storage *** this work,lead free relaxor ferroelectric ceramics in (1-x) Bi_(0.51)Na_(0.47)TiO_(3-x...
详细信息
Dielectric ceramics with high energy storage density and energy efficiency play an important role in high power energy storage *** this work,lead free relaxor ferroelectric ceramics in (1-x) Bi_(0.51)Na_(0.47)TiO_(3-x)Ba(Zr_(0.3)Ti_(0.7))O_(3)(BNT-BZT100x:x=0.20,0.30,0.40 and 0.50)system are fabricated by conventional solid-state sintering *** BNT-BZT100x ceramics are sintered dense with minimal pores,exhibiting pseudocubic symmetry and strong relaxor characteristic.A high energy storage density of 3.1 J/cm^(3) and high energy efficiency of 91% are simultaneously achieved in BNT-BZT40 ceramic with 0.1mm in thickness,at the applied electric field of 280 kV/*** temperature stability of the energy density is studied over temperature range of 20-160℃ ,showing minimal variation below 1.5%,together with the excellent cycling reliability(the variations of both energy density and efficiency are below 3% up to 106 cycles),making BNT-BZT40 ceramic promising candidate for high-temperature dielectric and energy storage applications.
Sometimes it is not enough for a DNN to produce an outcome. For example, in applications such as healthcare, users need to understand the rationale of the decisions. Therefore, it is imperative to develop algorithms t...
详细信息
How to identify high-potential talent (HIPO) earlier in their career always has strategic importance for human resource management. While tremendous efforts have been made in this direction, most existing approaches a...
详细信息
How to identify high-potential talent (HIPO) earlier in their career always has strategic importance for human resource management. While tremendous efforts have been made in this direction, most existing approaches are still based on the subjective selection of human resource experts. This could lead to unintentional bias and inconsistencies. To this end, in this paper, we propose a neural network based dynamic social profiling approach for quantitatively identifying HIPOs from the newly-enrolled employees by modeling the dynamics of their behaviors in organizational social networks. A basic assumption is that HIPOs usually perform more actively and have higher competencies than their peers to accumulate their social capitals during their daily work practice. Along this line, we first propose to model the social profiles of employees with both Graph Convolutional Network (GCN) and social centrality analysis in a comprehensive way. Then, an adaptive Long Short Term Memory (LSTM) network with global attention mechanism is designed to capture the profile dynamics of employees in the organizational social networks during their early career. Finally, extensive experiments on real-world data clearly validate the effectiveness of our approach as well as the interpretability of our results.
Disulfide bonds are vital for protein functions, but locating the linkage sites has been a challenge in protein chemistry, especially when the quantity of a sample is small or the complexity is high. In 2015,our labor...
详细信息
Disulfide bonds are vital for protein functions, but locating the linkage sites has been a challenge in protein chemistry, especially when the quantity of a sample is small or the complexity is high. In 2015,our laboratory developed a sensitive and efficient method for mapping protein disulfide bonds from simple or complex samples(Lu et al. in Nat Methods 12:329, 2015). This method is based on liquid chromatography–mass spectrometry(LC–MS) and a powerful data analysis software tool named p *** facilitate application of this method, we present step-by-step disulfide mapping protocols for three types of samples—purified proteins in solution, proteins in SDS-PAGE gels, and complex protein mixtures in solution. The minimum amount of protein required for this method can be as low as several hundred nanograms for purified proteins, or tens of micrograms for a mixture of hundreds of *** entire workflow—from sample preparation to LC–MS and data analysis—is described in great detail. We believe that this protocol can be easily implemented in any laboratory with access to a fastscanning, high-resolution, and accurate-mass LC–MS system.
暂无评论