Jacques J. Vidal

Implementing neural nets with programmable logic

Networks of Boolean programmable logic modules are presented as one purely digital class of artificial neural nets. The approach contrasts with the continuous analog framework usually suggested. Programmable logic networks are capable of handling many neural net applications. They avoid some of the limitations of threshold logic networks and present distinct opportunities. The network nodes are called dynamically programmable logic modules. They can be implemented with digitally controlled demultiplexers. Each node performs a Boolean function of its inputs which can be dynamically assigned. The overall network is therefore a combinational circuit and its outputs are Boolean global functions of the networks input variables. The approach offers definite advantages for VLSI implementation, namely, a regular architecture with limited connectivity, simplicity of the control machinery, natural modularity, and the support of a mature technology. >

Perturbation effects on stability of gravity receptors

Stability properties of gravity receptor transduction mechanisms were investigated by recording action potentials from the eighth nerve of cats maintained at two spatial positions both with and without an added click perturbation. The cells were demonstrated to present directional sensitivity, multivaluedness and wandering of mean rate trajectories. To the wandering were fitted appropriate stability boundaries corresponding to recent theories of finite time stability. Results supported the hypothesis that physiological stimuli result in local deformations of a flexible trampoline-like macula.

Process control with adaptive range coding

Dynamical control with adaptive range coding eliminates fundamental shortcomings found in earlier applications of range (course) coding which used fixed partitioning. Adaptive range coding has the advantages of efficient implementation, execution and generalization. With the adaptive algorithm, region shapes are continually adjusted during operation and will self-organize to reflect the global dynamics of the system and the environment. The system progressively develops a control map relating environmental states, control actions, and future reinforcements.

Machine operant conditioning

This research investigates learning of machine reflexes by applying punishment and reward reinforcement to teach artificial neuronlike systems a prescribed behavior. Stochastic neuronlike elements based on the classical weighted sum of inputs and threshold model can learn stimulus-response associations by emulated classical Pavlovian conditioning, i.e. make associations between conditioned and unconditioned stimuli and later responses. Several mathematical models have been developed which apply abstractions of classical conditioning to such threshold logic devices. Temporal sequences of stimulus-response associations can be dynamically learned by using operant conditioning when only aggregate external reinforcement is available.<<ETX>>

IMAGE RECOGNITION AND RECONSTRUCTION USING ASSOCIATIVE MAGNETIC PROCESSING

This paper presents a technique for image recognition, reconstruction, and processing using a novel massively parallel system. This device is a physical implementation of a Boltzmann machine type of neural network based on the use of magnetic thin films and opto-magnetic control. Images or patterns in the form of pixel arrays are imposed on the magnetic film using a laser in an external magnetic field. These images are learned and can be recalled later when a similar image is presented. A stored image is recallable even when a partial, noisy, or corrupted version of that image is imposed on the film. The system can also be used for feature detection and image compression. The operation and construction of the physical system is described, together with a discussion of the physical basis for its operation. The authors have developed Monte Carlo style computer simulations of the system for a variety of platforms, including serial workstations and hypercube configured parallel systems. They describe here some of the factors involved in computer simulations of the system, which can be fast and relatively simple in implementation. Simulation results are presented and, in particular, the behavior of the model under simulated annealing in the light of statistical physics is discussed. The simulation itself can be used as a neural network model capable of the functions ascribed to the physical device.

Generation of ECG prototype waveforms by piecewise correlational averaging

Describes a novel approach to the processing of ECG data. The procedure constructs a prototype waveform and generates timing statistics on the waveform components within the beats. Individual segments, defined heuristically to separate the P, QRS, and T waves in each recorded waveform, are aligned by correlation and averaged independently. The individual segment averages are subsequently concatenated to form the prototype. The new prototype is shown to correlate better with the ensemble of data waveforms than does the conventional time-locked average. The process also generates intrabeat statistics that describe the relative latency fluctuations within the waveform population.

An associative magnetic processor for image reconstruction

A novel, massively parallel system is presented for image memorization, processing, and reconstruction based on the use of magnetic thin films and optomagnetic control. Images in the form of pixel arrays are imposed on the film by locally heating it in the regions of high image intensity, using a laser with an external magnetic field. One of these stored images is recalled when a partial or corrupted version of that image is imposed on the film. The system can also be used for feature detection and image compression.<<ETX>>

Simulation of field problems using general-purpose

the university to organize an emerging analog computing laboratory within the Applied Physics Department. His early work was centered around direct analog method for solving partial differential equations with emphasis on heat transfer problems. Still associated with the University of Liege, he spent a year studying nuclear engineering at the French Atomic Center in Saclay and started the thesis that led him to a PhD degree from the Sorbonne, Paris, in 1961. Appointed lecturer in Liege, he taught the first course on analog computation officially offered by that university. In 1962, he made a trip to the U.S. and joined the faculty at UCLA as Assistant Professor in 1963. At

A unified approach to deterministic and random errors in hybrid loops

Hybrid computation in general owes its growing acceptance to the increasing number of sophisticated problems that neither digital nor analog computers alone can handle adequately. In most cases this apparently cumbersome approach is justified by the speed of analog computers, a speed impossible to attain in all-digital systems, even where the problem requires the memory and logic capabilities that only digital computers can provide.