Counterpropagation networks applied to the classification of alkanes through infrared spectra

Enhanced uni-flow counterpropagation networks are used as pattern recognition systems and applied to the identification of chemical structure from corresponding infrared spectra. It is shown that such networks are more suitable for this type of problem than backpropagation networks, both in terms of training times and network performance. The problem of optimum classification between highly similar infrared spectra is addressed, and factors such as training set size, sampling rate, data pre-processing, output data representation and the number of Kohonen layer nodes are considered in this context. It is shown that such networks may achieve rates of correct classification in excess of 90%, although the learning of correct decision boundaries is highly sensitive to the above parameters in cases where the non-informational content of training and test data varies considerably with respect to the informational content, and hence clustering of classes in pattern space is incomplete.     

Effect of dry etching on surface properties of III-nitrides

Dry etched InAlN and GaN surfaces have been characterized by current-voltage measurement, Auger electron spectroscopy, and atomic force microscopy. Electron cyclotron resonance discharges of BCl₃. BCl₃/Ar, BCl₃/N₂, or BCl₃/N₂ plus wet chemical etch all produce nitrogen surfaces that promote leakage current in rectifying gate contacts, with the BCl₃/N₂ plus wet chemical etch producing the least disruption on the surface properties. The conductivity of the immediate InAlN or GaN surface can be increased by preferential loss of N during BCl₃ plasma etching, leading to poor rectifying contact characteristics when the gate metal is deposited on this etched surface. Careful control of plasma chemistry, ion energy, and stoichiometry of the etched surface are necessary for acceptable pinch-off characteristics. Hydrogen passivation during the etch was also studied.     

Isolation of cell cycle-dependent gamma ray-sensitive Chinese hamster ovary cell

A technique for the isolation of gamma ray-sensitive Chinese hamster ovary (CHO) cell mutants is described, which uses nylon cloth replica plating and photography with dark-field illumination to directly monitor colonies for growth after gamma irradiation. Two gamma ray-sensitive mutants were isolated using this method. One of these cells (XR-1) had a two-slope survival curve: an initial steep slope and then a flattening of the curve at about 10% survival. Subsequently, it was found that this cell is sensitive to gamma irradiation in G1, early S, and late G2 phases of the cell cycle, whereas in the resistant phase (late S phase) its survival approaches that of the parental cells. The D37 in the sensitive G1 period is approximately 30 rads, compared with 300 rads of the parental cell. This mutant cell is also sensitive to killing by the DNA breaking agent, bleomycin, but is relatively insensitive to UV light and ethyl methane sulfonate, suggesting that the defect is specific for agents that produce DNA strand breakage.     

Hardware implementation of a `wired-once' neural net in thin-filmtechnology on a glass substrate

To prove the feasibility of implementing artificial neural networks on large inexpensive substrates, a net designed and fabricated on a glass wafer using hydrogenated-amorphous-silicon-based technology (a-Si:H) is discussed. The net functions as an autoassociative memory in which binary numbers corresponding to 28, 56, 112, and 224 are stored. Learning of the weight matrix is carried out with the associative memory algorithm using the delta rule. Phosphorus-doped microcrystalline silicon with a resistivity of 100 to 300 Ω-cm was used for the fabrication of the weight (synapse) resistors. Inverters with a beta of one were used to form negative-weight synapses, and inverters with a beta of 10 were used for the thresholding elements (neurons). The net functions surprisingly well; it filters both the learned numbers and some numbers of the form N=4k (with k an integer), and maps other random numbers to the closest one accepted, even though the experimental weight matrix is not identical to the theoretical one     

Coronary Artery Disease?Correlates Between Diastolic Auditory Characteristics and Coronary Artery Stenoses

A noninvasive approach to detecting coronary artery disease analyzes thoracic sounds to isolate acoustical correlates of stenosed coronary arteries., The analysis includes time windowing, frequency (power spectra) windowing, and averaging of thoracic sounds from normal and diseased patients. Initial results indicate that an above normal percentage of high-frequency (180-300 Hz) energy is closely associated with narrowed coronary arteries.     

A simple high-temperature infrared cell and its application to the dehydroxylation of kaolinite

A simple method is described for making infrared measurements on solids at temperatures up to the melting point of KBr by using a fine nichrome heating coil wound directly on to the KBr disc containing the sample either in suspension or sedimented on to the surface. The usefulness of the method is illustrated by intensity measurements of the hydroxyl stretching bands in kaolinite as a function of temperature, which are compared with results obtained by the more conventional quenching method.     

Pseudotumour of the lung caused by infection with Bacillus sphaericus

A patient is described who suffered from severe chronic asthma complicated by repeated chest infections. She developed a large gelatinous pseudotumour of the lung which was found to be caused by Bacillus sphaericus.