Continuous speech recognition based on ICA and geometrical learning

We investigate the use of independent component analysis (ICA) for speech feature extraction in digits speech recognition systems. We observe that this may be true for recognition tasks based on Geometrical Learning with little training data. In contrast to image processing, phase information is not essential for digits speech recognition. We therefore propose a new scheme that shows how the phase sensitivity can be removed by using an analytical description of the ICA-adapted basis functions. Furthermore, since the basis functions are not shift invariant, we extend the method to include a frequency-based ICA stage that removes redundant time shift information. The digits speech recognition results show promising accuracy. Experiments show that the method based on ICA and Geometrical Learning outperforms HMM in a different number of training samples.

Research on multi-degree-of-freedom neurons with weighted graphs

In this paper, we redefine the sample points set in the feature space from the point of view of weighted graph and propose a new covering model - Multi-Degree-of-Freedorn Neurons (MDFN). Base on this model, we describe a geometric learning algorithm with 3-degree-of-freedom neurons. It identifies the sample points secs topological character in the feature space, which is different from the traditional "separation" method. Experiment results demonstrates the general superiority of this algorithm over the traditional PCA+NN algorithm in terms of efficiency and accuracy.

A novel image restoration algorithm based on high-dimensional space geometry

A novel image restoration approach based on high-dimensional space geometry is proposed, which is quite different from the existing traditional image restoration techniques. It is based on the homeomorphisms and "Principle of Homology Continuity" (PHC), an image is mapped to a point in high-dimensional space. Begin with the original blurred image, we get two further blurred images, then the restored image can be obtained through the regressive curve derived from the three points which are mapped form the images. Experiments have proved the availability of this "blurred-blurred-restored" algorithm, and the comparison with the classical Wiener Filter approach is presented in final.

Continuous speech research based on HyperSausage Neuron

In this paper, we presents HyperSausage Neuron based on the High-Dimension Space(HDS), and proposes a new algorithm for speaker independent continuous digit speech recognition. At last, compared to HMM-based method, the recognition rate of HyperSausage Neuron method is higher than that of in HMM-based method.

Sub-wavelength beam lasers with surface plasmon structures

The simulation of a plasmonic very-small-aperture laser is demonstrated in this paper. It is an integration of the surface plasmon structure and very-small-aperture laser (VSAL). The numerical results demonstrate that the transmission field can be confined to a spot with subwavelength width in the far field (3.5 mu m far from the emitting surface), and the output power density can be enhanced over 30 times of the normal VSAL. Such a device can be useful in the application of a high resolution far-field scanning optical microscope.

Simultaneous multiwavelength observations of sagittarius A

We observed Sgr A* using the Very Large Array (VLA) and the Giant Metrewave Radio Telescope (GMRT) at multiple centimeter and millimeter wavelengths on 2003 June 17. The measured flux densities of Sgr A*, together with those obtained from the Submillimeter Array (SMA) and the Keck II 10 m telescope on the same date, are used to construct a simultaneous spectrum of Sgr A* from 90 cm to 3.8 mu m. The simultaneous spectrum shows a spectral break at about 3.6 cm, a possible signature of synchrotron self-absorption of the strong radio outburst that occurred near epoch 2003 July 17. At 90 cm, the flux density of Sgr A* is 0.22 +/- 0.06 Jy, suggesting a sharp decrease in flux density at wavelengths longer than 47 cm. The spectrum at long cm wavelengths appears to be consistent with free-free absorption by a screen of ionized gas with a cutoff similar to 100 cm. This cutoff wavelength appears to be three times longer than that of similar to 30 cm suggested by Davies, Walsh, & Booth based on observations in 1974 and 1975. Our analysis suggests that the flux densities of Sgr A* at wavelengths longer than 30 cm could be attenuated and modulated by stellar winds from massive stars close to Sgr A*.

不同培肥措施下种植制度及撂荒对土壤微生物量碳氮的影响

以黄土高原南部半湿润易旱区已进行17年的田间定位试验为研究对象,研究了不同培肥措施(不施肥、施用氮磷钾及氮磷钾与有机肥配合施用)下两种种植制度(一年1熟及一年两熟)和撂荒对土壤微生物量碳、氮(SMBC、SMBN)及可溶性有机碳、氮(SOC、SON)等含量的影响。结果表明,与一年1熟的小麦-休闲种植制度相比,一年两熟小麦-玉米轮作提高了0~10cm土层SMBC、SMBN、有机碳(TOC)、全氮(TN)和土壤SOC、SON的含量,而对10~20cm土层上述测定指标影响不大。与不施肥(CK)或单施化肥处理(NPK)下小麦-休闲和小麦-玉米轮作方式相比,撂荒处理显著提高了0~10cm土层各测定指标的含量。不同培肥措施相比,氮磷钾配施有机肥显著提高了0~10cm、10~20cm土层SMBC、SMBN含量;NPK处理0~10cm土层SMBN含量显著增加,10~20cm土层SMBN和0~10cm、10~20cm土层SMBC含量增加但未达显著水平。不同培肥措施和种植制度对SMBC/TOC和SMBN/TN的比例无明显影响。

STUDY ON PREPARATION OF GAN AND COSI2 EPITAXIAL-FILMS BY MASS ANALYZED LOW-ENERGY DUAL ION-BEAM EPITAXY

The Mass Analyzed Low Energy Dual Ion Beam Epitaxy (MALE-DIBE) system has been designed and constructed in our laboratory. We believe that the system, which was installed and came into full operation in 1988, is the first facility of this kind. With our system we have carried out studies, for the first time, on compound synthesis of GaN and CoSi2 epitaxial thin films. RHEED and AES results show that GaN films, which were deposited on Si and sapphire substrates, are monocrystalline and of good stoichiometry. To our knowledge, GaN film heteroepitaxially grown on Si. which is more lattice-mismatched than GaN on sapphire, has not been reported before by other authors. RBS and TEM investigations indicated a rather good crystallinity of CoSi2 with a distinct interface between CoSi2 and the Si substrate. The channelling minimum yield chi(min) from the Co profile is approximately 4%. The results showed that the DIBE system with simultaneous arrival of two beams at the target is particularly useful in the formation of novel compounds at a relatively low substrate temperature.

DIRECT ION-BEAM DEPOSITION OF CARBON-FILMS ON SILICON IN THE ION ENERGY-RANGE OF 15-500 EV

Direct ion beam deposition of carbon films on silicon in the ion energy range of 15-500 eV and temperature range of 25-800-degrees-C has been studied. The work was carried out using mass-separated C+ and CH3+ ions under ultrahigh vacuum. The films were characterized with x-ray photoelectron spectroscopy, Raman spectroscopy, transmission electron microscopy, and transmission electron diffraction analysis. In the initial stage of the deposition, carbon implanted into silicon induced the formation of silicon carbide, even at room temperature. Further carbon ion bombardment then led to the formation of a carbon film. The film properties were sensitive to the deposition temperature but not to the ion energy. Films deposited at room temperature consisted mainly of amorphous carbon. Deposition at a higher temperature, or post-deposition annealing, led to the formation of microcrystalline graphite. A deposition temperature above 800-degrees-C favored the formation of microcrystalline graphite with a preferred orientation in the (0001) direction. No evidence of diamond formation in these films was observed.