Abstract-Over the past few decades, a large family of algorithms-supervised or unsupervised; stemming from statistics or geometry theory-has been designed to provide different solutions to the problem of dimensionality reduction. Despite the different motivations of these algorithms, we present in this paper a general formulation known as graph embedding to unify them within a common framework. In graph embedding, each algorithm can be considered as the direct graph embedding or its linear/kernel/tensor extension of a specific intrinsic graph that describes certain desired statistical or geometric properties of a data set, with constraints from scale normalization or a penalty graph that characterizes a statistical or geometric property that should be avoided. Furthermore, the graph embedding framework can be used as a general platform for developing new dimensionality reduction algorithms. By utilizing this framework as a tool, we propose a new supervised dimensionality reduction algorithm called Marginal Fisher Analysis in which the intrinsic graph characterizes the intraclass compactness and connects each data point with its neighboring points of the same class, while the penalty graph connects the marginal points and characterizes the interclass separability. We show that MFA effectively overcomes the limitations of the traditional Linear Discriminant Analysis algorithm due to data distribution assumptions and available projection directions. Real face recognition experiments show the superiority of our proposed MFA in comparison to LDA, also for corresponding kernel and tensor extensions.

This paper attempts to provide some new understanding of the mechanical as well as thermal effects of the Tibetan Plateau (TP) on the circulation and climate in Asia through diagnosis and numerical experiments. The air column over the TP descends in winter and ascends in summer and regulates the surface Asian monsoon flow. Sensible heating on the sloping lateral surfaces appears from the authors’ experiments to be the major driving source. The retarding and deflecting effects of the TP in winter generate an asymmetric dipole zonal-deviation circulation, with a large anticyclone gyre to the north and a cyclonic gyre to the south. Such a dipole deviation circulation enhances the cold outbreaks from the north over East Asia, results in a dry climate in south Asia and a moist climate over the Indochina peninsula and south China, and forms the persistent rainfall in early spring (PRES) in south China. In summer the TP heating generates a cyclonic spiral zonal-deviation circulation in the lower troposphere, which converges toward and rises over the TP. It is shown that because the TP is located east of the Eurasian continent, in summertime the meridional winds and vertical motions forced by the Eurasian continental-scale heating and the TP local heating are in phase over the eastern and central parts of the continent. The monsoon in East Asia and the dry climate in middle Asia are therefore intensified.

Canonical correlation analysis (CCA) is applied to analyze the frequency components of steady-state visual evoked potentials (SSVEP) in electroencephalogram (EEG). The essence of this method is to extract a narrowband frequency component of SSVEP in EEG. A recognition approach is proposed based on the extracted frequency features for an SSVEP-based brain computer interface (BCI). Recognition Results of the approach were higher than those using a widely used fast Fourier transform (FFT)-based spectrum estimation method.

Background: Gold nanoparticle toxicity research is currently leading towards the in vivo experiment. Most toxicology data show that the surface chemistry and physical dimensions of gold nanoparticles play an important role in toxicity. Here, we present the in vivo toxicity of 5, 10, 30, and 60 nm PEG-coated gold nanoparticles in mice. Methods: Animal survival, weight, hematology, morphology, organ index, and biochemistry were characterized at a concentration of 4000 µg/kg over 28 days. Results:The PEG-coated gold particles did not cause an obvious decrease in body weight or appreciable toxicity even after their breakdown in vivo. Biodistribution results show that 5 nm and 10 nm particles accumulated in the liver and that 30 nm particles accumulated in the spleen, while the 60 nm particles did not accumulate to an appreciable extent in either organ. Transmission electron microscopic observations showed that the 5, 10, 30, and 60 nm particles located in the blood and bone marrow cells, and that the 5 and 60 nm particles aggregated preferentially in the blood cells. The increase in spleen index and thymus index shows that the immune system can be affected by these small nanoparticles. The 10 nm gold particles induced an increase in white blood cells, while the 5 nm and 30 nm particles induced a decrease in white blood cells and red blood cells. The biochemistry results show that the 10 nm and 60 nm PEG-coated gold nanoparticles caused a significant increase in alanine transaminase and aspartate transaminase levels, indicating slight damage to the liver. Conclusion: The toxicity of PEG-coated gold particles is complex, and it cannot be concluded that the smaller particles have greater toxicity. The toxicity of the 10 nm and 60 nm particles was obviously higher than that of the 5 nm and 30 nm particles. The metabolism of these particles and protection of the liver will be more important issues for medical applications of gold-based nanomaterials in future.

Garnet peridotites occur as lenses, blocks or layers within granulite–amphibolite facies gneiss in the Dabie‐Sulu ultra‐high‐pressure (UHP) terrane and contain coesite‐bearing eclogite. Two distinct types of garnet peridotite were identified based on mode of occurrence and petrochemical characteristics. Type A mantle‐derived peridotites originated from either: (1) the mantle wedge above a subduction zone, (2) the footwall mantle of the subducted slab, or (3) were ancient mantle fragments emplaced at crustal depths prior to UHP metamorphism, whereas type B crustal peridotite and pyroxenite are a portion of mafic–ultramafic complexes that were intruded into the continental crust as magmas prior to subduction. Most type A peridotites were derived from a depleted mantle and exhibit petrochemical characteristics of mantle rocks; however, Sr and Nd isotope compositions of some peridotites have been modified by crustal contamination during subduction and/or exhumation. Type B peridotite and pyroxenite show cumulate structure, and some have experienced crustal metasomatism and contamination documented by high 87Sr/86Sr ratios (0.707–0.708), low εNd(t) values (−6 to −9) and low δ18O values of minerals (+2.92 to +4.52). Garnet peridotites of both types experienced multi‐stage recrystallization; some of them record prograde histories. High‐P–T  estimates (760–970 °C and 4.0–6.5±0.2 GPa) of peak metamorphism indicate that both mantle‐derived and crustal ultramafic rocks were subducted to profound depths >100 km (the deepest may be ≥180–200 km) and experienced UHP metamorphism in a subduction zone with an extremely low geothermal gradient of <5 °C km−1.