Neural nets

Meer over het boek

This comprehensive collection covers a wide array of topics in neural networks, data visualization, and pattern recognition. It begins with an invited paper discussing latent variables and topographic mappings, followed by review papers on fuzzy neural networks and gradient calculations in locally recurrent networks. The Eduardo R. Caianiello Lecture focuses on a progressive learning neural network for identification and control. Mathematical models explore various aspects, including testing sizes, entropy comparisons for classification, and Hebbian models for spatial representation. The section also addresses discrete Hopfield networks for optimization problems and the approximation rates of multivariable functions using one-hidden-layer networks. Additionally, it examines energy functionals, VC-dimension performance bounds, and attractor neural networks as semantic memory models. In pattern recognition and robotics, topics include image reconstruction with RBF networks, MLP-based character spotting, video data compression, and speech recognition applications. The section also covers periodicity analysis and mobile robot navigation using visual landmarks. Applications range from ecological data analysis and hybrid fuzzy-neural systems in plasma engineering to railway anti-skating systems and meteorological visibility forecasting. The collection also features self-organization in large medical databases and software integra