Neural processing is applied to a particle discriminator problem in a high-energy experimental physics environment. Using the energy deposition profile for incoming particles provided by an energy measurement detector with high granularity (a calorimeter), a two-layered neural network discriminator is trained on experimental data to identify electrons, pions and muons. During the training phase, the neural network discriminator is able to identify impurities in the original data sample, and

more »

... sical physics methods do validate such outsider identification. After removing these impurities from the data sample, the neural network was retrained and was able to identify electrons, pions and muons with efficiencies of 100%, 99% and 100%, respectively. The discriminator may be implemented on fast digital signal processor technology for on-line operation.