Integration of prior knowledge of measurement noise in kernel density classification

Samples can be measured with different precisions and reliabilities in different experiments, or even within the same experiment. These varying levels of measurement noise may deteriorate the performance of a pattern recognition system, if not treated with care. Here we seek to investigate the benefit of incorporating prior knowledge about measurement noise into system construction. We propose a kernel density classifier which integrates such prior knowledge. Instead of using an identical kernel for each sample, we transform the prior knowledge into a distinct kernel for each sample. The integration procedure is straightforward and easy to interpret. In addition, we show how to estimate the diverse measurement noise levels in a real world dataset. Compared to the basic methods, the new kernel density classifier can give a significantly better classification performance. As expected, this improvement is more obvious for small sample size datasets and large number of features.     

Secondary Aluminum Production Waste — A Raw Material for the Refractory Industry

Secondary aluminum waste products, and processes that occur during heating and mineral formation on introducing fine periclase powder into the composition of waste products are studied. It is demonstrated that secondary aluminum production waste may be used as a raw material for synthesizing aluminomagnesia spinels. Translated from Novye Ogneupory, No. 6, pp. 15–18, June 2008.     

Testing a new composition of periclase spinel objects

On the basis of periclase powder (70%) and aluminomagnesia spinel (30%), synthesized from secondary aluminum production waste and caustic magnesite, test objects are produced of periclase spinel composition satisfying the requirements of GOST 1579-006-00190495-98. The articles are successfully tested in the working layer for lining tunnel furnace cars of OAO Kombinat Magnezit. __________ Translated from Novye Ogneupory, No. 8, pp. 16–17, August 2008.     

Synthesis of nanocrystalline solid solutions based on zirconia and hafnia

Nanosized (2–8 nm) amorphous powders of the solid solution based on zirconia and hafnia are synthesized through back coprecipitation upon treatment of gels at temperatures from +20 to −77°C. Heat treatment of these powders at temperatures up to 1000 and above 1100°C leads to the formation of cubic (fluorite type, O _h ⁵ = Fm3m) and tetragonal phases of the Zr₈₂Hf₁₀Y₃Ce₅O _x composition, respectively. It is revealed that a decrease in the synthesis temperature (from +20°C to −6°C) results in a decrease in the size of gel agglomerates from 30 to 1 μm. Recrystallization processes in the gels prepared using cryochemical treatment are developed very slowly in the temperature range 500–1200°C (the crystallite size does not exceed 25 nm). Original Russian Text ? T.I. Panova, V.B. Glushkova, A.E. Lapshin, 2008, published in Fizika i Khimiya Stekla.     

Aggregate stability of ZrO2 aqueous sols in electrolyte solutions

The aggregate stability of submicron and nanosized ZrO₂ aqueous sols of different origins and different dispersities at pH 3–10 in the KCl concentration range 10^?3–10^?1 M is investigated by flow ultramicroscopy and photometry. The results obtained are analyzed in the framework of the extended Derjaguin-Landau-Verwey-Overbeek theory and the Muller-Martynov theory of reversible aggregation. The extension of boundary layers of water near the surface of the ZrO₂ particles is estimated.     

Influence of protein and lipid concentration of the test lubricant on the wear of ultra high molecular weight polyethylene

To gain a better understanding of the ultra-high molecular weight polyethylene (UHMWPE) wear mechanism in the physiological environment, the effects of protein and lipid constituents of synovial fluid on the specific wear rate of UHMWPE were examined experimentally. The multidirectional sliding pin-on-plate wear tester was employed to simulate the simplified sliding condition of hip joint prostheses. Bovine serum γ-globulin and synthetic l--DPPC were used as model protein and lipid constituents of synovia, respectively. Results of the wear test indicated that the UHMWPE wear rate primarily depended on the protein concentration of the test lubricant. Lipids acted as a boundary lubricant and reduced polyethylene wear in the low protein lubricants. However, the polyethylene wear rate increased with increasing lipid concentrations if the protein concentration was within the physiological level. Increased interactions between protein and lipid molecules and lipid diffusion to polyethylene surface might be responsible for the increased wear.     

Perfect Blind-Channel Shortening for Multicarrier Systems

In multicarrier systems, when the order of a channel impulse response is larger than the length of the cyclic prefix (CP), there is a significant performance degradation due to interblock interference (IBI). This paper proposes a blind-channel shortening method in which the equalizer parameter vector is formed by the noise subspace of the received signal correlation matrix so that the output power is maximized. The proposed method can not only shorten the effective channel impulse response to within the CP length but also maximize the output signal-to-interference-and-noise ratio while eliminating the IBI. We point out that the performance depends on the choice of a decision delay and propose a simple method for determining the appropriate delay. We propose both a batch algorithm and an adaptive algorithm and show by simulation that they are superior to the conventional algorithms.     

Design Techniques of CMOS Ultra-Wide-Band Amplifiers for Multistandard Communications

This paper presents design techniques of CMOS ultra-wide-band (UWB) amplifiers for multistandard communications. The goal of this paper is to propose a compact, simple, and robust topology for UWB low-noise amplifiers, which yet consumes a relatively low power. To achieve this goal, a common-gate amplifier topology with a local feedback is employed. The first amplifier uses a simple inductive peaking technique for bandwidth extension, while the second design utilizes a two-stage approach with an added gain control feature. Both amplifiers achieve a flat bandwidth of more than 6 GHz and a gain of higher than 10 dB with supply voltages of 1.8-2.5 V. Designs with different metal thicknesses are compared. The advantage of using thick-metal inductors in UWB applications depends on the chosen topology.     

A Low-Supply-Voltage CMOS Sub-Bandgap Reference

A low-power (21 $muhbox{W}$ ) bandgap reference source that is operable from a nominal supply voltage of 1.4 V is described. The circuit provides an output voltage equal to the bandgap voltage having a low output resistance and allows resistive loading. It does not use resistors or operational amplifiers. Thus, the design is suitable for fabrication in any digital CMOS technology. The circuit uses a current conveyor and current mirrors to convert the proportional to absolute temperature voltage into a current using a MOSFET. The current is converted back to a voltage by using the functional inverse of the FET $v-i$ characteristics. This makes the voltage gain linear and temperature independent. The absence of back-gate bias is the reason for achieving the low supply voltage of operation. Simulation results using the transistor models for the 0.18-$mu$m TSMC process show that the voltage-variation over the temperature range 0 to 100 $^{circ} {hbox {C}}$ is $≪$1 mV.