Abstruct-Implementation of the adaptive resonance theory (ART) of neural networks has been a thorny problem for several years. This work presents a novel solution to the problem by using an optical correlator, allowing the large body of correlator research to be leveraged in the implementation of ART. The implementation takes advantage of the fact that one ART-based architecture, known as ART1, can be broken into several parts, some of which are better to implement in parallel. The control structure of ART, often regarded as its most complex part, is actually not very time consuming and can be done in electronics. The bottom-up and top-down gated pathways, however, are very time consuming to simulate and are difficult to implement directly in electronics due to the high number of interconnections. Tbo face simplify this. The first is that the pathways are computing a set of inner products. These inner products represent as least 80% of the computation time of the ARTl implementation. The second insight, our contribution, is that implementing the inner products optically, and the rest of the network in electronics, is a very effective marriage of the two technologies to realize the ARTl network. In addition to the design, we present experiments with a laboratory prototype to illustrate its feasibility and to discuss implementation details that arise in practice. This device potentially can significantly outperform alternative implementations of ARTl by as much as two to three orders of magnitude in problems requiring especially large input fields. It should be noted that all of these results apply to just one of the various ART architectures, known as ART1, but that other ART networks and other neural nets in general also use inner products and could benefit from this work as well.

A numerical model for optically triggered switching in bulk GaAs is presented. First, a one-dimensional model is described and calculated behavior compared to experimental observations. Results from the one-dimensional model are not consistent with observed switching behavior. The model is then modified to include filament formation. Results from the modified model agree qualitatively and quantitatively with experimental data. Details of the dynamic behavior of the device are shown and a unified picture of the switching phenomenon presented. On the basis of the agreement of the numerical model and experimental observation it is concluded that switching is a result of localized impact ionization creating a conductive filament channel through the bulk material.

Ultrashort-pulse (femtosecond-duration) two-photon laser-induced fluorescence (fs-TPLIF) of an inert gas tracer krypton (Kr) is investigated. A detailed spectroscopic study of fluorescence channels followed by the 5p^'←←4p excitation of Kr at 204.1 nm is reported. The experimental line positions in the 750-840 nm emission region agree well with the NIST Atomic Spectra Database. The present work provides an accurate listing of relative line strengths in this spectral region. In the range of laser pulse energies investigated, a quadratic dependence was observed between the Kr-TPLIF signal and the laser pulse energy. The single-laser-shot 2D TPLIF images recorded in an unsteady jet demonstrate the potential of using fs excitation at 204.1 nm for mixing and flow diagnostic studies using Kr as an inert gas tracer.

An analytic method for inversion problems of the Fredholm integral type is developed. The application of this formalism to error analysis and optimal experimental design is discussed. Finally, results from this method for a simulated experiment are compared with other commonly used inversion techniques.

There has been difficulty in achieving a fully parallel, digital optical adder or multiplier. The primary obstacle is the carry operation inherent in any fixed-radix number system. The concepts of residue number representation and symbolic substitution can be combined to produce a parallel optical arithmetic/logic unit.