A Neuro-fuzzy Coding for Processing Incomplete Data: Application to the Classification of Seismic Events

This letter presents a method for modelling and processing incomplete data in connectionist systems. The approach consists in applying a neuro-fuzzy coding to the input data of a neural network. After an introduction to the different kinds of imperfections, we propose a neuro-fuzzy coding in order to take incomplete data into account. We show the efficiency of this coding on the problem of the classification of seismic events. The results show that a neuro-fuzzy coding of the inputs of a neural network increases the performance and classifies incomplete data with little affect on the results.     

Molecular species of phosphoglycerides in liver microsomes of rats fed a fat-free diet

The influence of a fat-free diet on the molecular species composition of phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylinositol (PI) of rat liver microsomes was studied by using reversed-phase high-pressure liquid chromatography. In the three phosphoglyceride classes analyzed, the fat-free diet produced a large decrease in the 18:0/20:4n-6 species but less important changes were found in the 16:0/20:4n-6 species. In PC, the most abundant phosphoglyceride class of rat liver microsomes, the fall in the 18:0/20:4n-6 species was counterbalanced mainly by an enhancement in the 16:0/18:1n-9 species although it was not evident in PE. In PI, the decrease in the 18:0/20:4n-6 species was counterbalanced by an increase in the 18:0/20:3n-9 species. Fluorescence polarization measurements of 1,7-diphenyl-1,3,5-hexatriene in liposomes of 16:0/18:1n-9, 18:0/18:1n-9-, 16:0/20:4n-6-, and 18:0/20:4n-6-PC indicated that the change in the saturated fatty acid in the sn-1 position accompanying the replacement of 20:4n-6 by 18:1n-9 could be very important for a homeoviscous compensation, maintaining the membrane physical properties without large alterations in spite of the essential fatty acid deficiency due to the fat-free diet.     

Polyurethane methacrylate/silicone interpenetrating polymer networks synthesis,thermal and mechanical properties

The synthesis of an interpenetrating polymer network (IPN) combining a polyurethane methacrylate network (PUMA) and a silicone network is reported. The PUMA network is synthesized by UV-light cure. The silicone network is formed through a condensation between α, ω dihydroxy polydimethyl siloxane and γ-methacryloxypropyl trimethoxy silane (γ-MPS) as a cross-linking agent. The IPN is prepared by different mechanism: radical and condensation types. According to thermogravimetric analysis of the hybrid material, the thermal stability stayed unchanged but the kinetic of degradation changed. Tg decreased with increasing silica content. The thermal cure process under humid atmosphere influence properties just for PUMA/4.2%Si^UV+T. Condensation between γ-MPS decreases the penetration depth from 158 to 82 μm and increases the mechanical glass transition temperature from 106 to 141 °C.     

Two analog counters for neural network implementation

Simple analog circuits which are useful for the implementation of the synchronous Boltzmann machine learning algorithms are presented. A simple charge-transfer-based analog counter is described. The authors give a functional model of its behavior and analyze the differences between this model and the counter implementation. They also present simulation results and the test of a prototype. Along the same lines, they study a switched-current-based counter, which achieves better results (dynamic range, linearity) through higher complexity     

Chip-in-the-loop learning algorithm for Boltzmann machine

A new algorithm is presented for chip-in-the-loop learning of the synchronous Boltzmann machine, which has two distinct advantages: it allows optimal use of analogue circuit dynamic range, unlike the conventional learning algorithm, and it features a coherent dual weight representation, which improves learning efficiency