Neural networks for contract bridge bidding

The objective of this study is to explore the possibility of capturing the reasoning process used in bidding a hand in a bridge game by an artificial neural network. We show that a multilayer feedforward neural network can be trained to learn to make an opening bid with a new hand. The game of bridge, like many other games used in artificial intelligence, can easily be represented in a machine. But, unlike most games used in artificial intelligence, bridge uses subtle reasoning over and above the agreed conventional system, to make a bid from the pattern of a given hand. Although it is difficult for a player to spell out the precise reasoning process he uses, we find that a neural network can indeed capture it. We demonstrate the results for the case of one-level opening bids, and discuss the need for a hierarchical architecture to deal with bids at all levels.     

Screen‐printed biosensor for allergens

Allergen levels in indoor environments, leading to many diseases, eg asthma, rhinitis and conjunctivitis, affect a large and increasing fraction of the population. A quite effective and inexpensive method of a rough but very rapid overall assessment of total allergen level in the environment has been developed. The method involved estimation of protein in allergen extracts by screen‐printed electrodes using two different techniques. The biosensor comprised a rhodinised carbon working electrode, a silver/silver chloride reference electrode and a carbon counter electrode. In the first method the enzyme protease reacted with allergen protein to release amino acid, which produced hydrogen peroxide in the presence of amino acid oxidase. This was detected amperometrically. The second method used potassium bromide as electrolyte and the electrode was subjected to dual potential. Bromine, released due to electrolysis at higher potential, was consumed by the allergen protein at lower potential. In the first method, a unique technique was used to microencapsulate the enzyme protease and immobilise it on the surface of the electrode by in‐situ polymerisation to avoid contact with the amino acid oxidase. A total of seven allergens were tested and the results gave a good correlation with the standard protein measurement method. Environmental specimens from indoors, schools and workplaces can be evaluated for the aeroallergens produced by dust mites, animal hairs, cockroach debris, pollens, etc as a means of determining the exposure risk. Copyright © 2005 Society of Chemical Industry     

Transient analysis of electromagnetic systems with multiple lumped nonlinear loads

A novel method for transient analysis of electromagnetic systems with multiple lumped nonlinear loads is presented. The uniqueness of this approach is that we develop time-domain Green's functions for the multiport linear part of the electromagnetic system by suitably terminating the ports. This ensures a short duration of Green's functions. Hence the amount of frequency-domain data necessary to obtain the time-domain Green's functions is modest. The application of this technique to an arbitrary excitation is just a straightforward convolution. With this technique one can analyze responses of systems with arbitrary nonlinear loads (even with memory) as we have at any time instant Thevenin's equivalent of the linear portion of the electromagnetic system. Examples are presented to illustrate the application of this technique to multiport nonlinearly loaded antennas.     

Impact of nonuniform graded dopant profile in polysilicon gate on gate leakage current

In a MOSFET, a nonuniform, graded vertical dopant profile in the polysilicon gate causes a potential drop at the polysilicon/oxide interface. In this paper, the effect of this potential drop on the gate leakage current has been evaluated for the first time. The extent of variations of this affected gate leakage current with gate oxide thickness, gate length, and gate and drain bias conditions have been assessed with device simulation for an nMOS at 0.13 /spl mu/m low-voltage process. The results provide a guideline to the severity of this effect from the point of view of device and circuit operation and standby power consumption.     

Time-domain measurements with the Hewlett-Packard network analyzerHP 8510 using the matrix pencil method

The generalized pencil of function (GPOF) method, also known as the matrix pencil method, is used to improve the resolution of HP 8510B network analyzer data in the time domain. This method provides for much higher resolution than the Fourier techniques. A comparison of the two methods is given for the example of the Beatty standard. The examples show that a parametric technique such as the GPOF can provide accurate and reliable results with a high degree of resolution even when the fast Fourier transform (FFT)-based technique fails     

A double-layer current conduction model for high-κ gate dielectric materials with interfacial oxide or silicate layer

The current-voltage (I-V) characteristics of metal-oxide-semiconductor (MOS) structures with hafnium oxide as the gate dielectric film were studied. Sharp shifts from a low-voltage ohmic regime to a tunneling conduction were observed in the high-voltage range. The paper demonstrates that this behavior can be described very well with a double-layer dielectric model. Excellent fittings of the experimental curves were obtained and the related key structural and physical parameters were obtained. The model fitting further suggests the optimal annealing conditions for preparing the hafnium oxide films.     

Semi-supervised model adaptation for statistical machine translation

Statistical machine translation systems are usually trained on large amounts of bilingual text (used to learn a translation model), and also large amounts of monolingual text in the target language (used to train a language model). In this article we explore the use of semi-supervised model adaptation methods for the effective use of monolingual data from the source language in order to improve translation quality. We propose several algorithms with this aim, and present the strengths and weaknesses of each one. We present detailed experimental evaluations on the French–English EuroParl data set and on data from the NIST Chinese–English large-data track. We show a significant improvement in translation quality on both tasks.