Artificial neural networks in pattern recognition

More about the book

The book covers a range of topics in machine learning, focusing on both unsupervised and supervised learning techniques. It begins with unsupervised learning methods, including effective nonparametric estimation of probability density functions and comparisons of spatio-temporal organization maps for speech recognition. It also discusses adaptive feedback inhibition to enhance pattern discrimination and various semi-supervised learning strategies. In the realm of supervised learning, the text explores training radial basis functions via gradient descent and presents a local tangent space alignment-based transductive classification algorithm. It highlights incremental manifold learning and introduces a convolutional neural network designed to tolerate synaptic faults, particularly for low-power analog hardware applications. Support vector learning is addressed through regression using Mahalanobis kernels and incremental training methods for support vector machines. The book also examines multiple classifier systems, including their application in embedded string patterns and facial feature localization using multiple neural networks. Visual object recognition is another key focus, detailing object detection with sparse convolutional neural networks and image classification through geometric appearance learning. Additionally, it touches on eye detection systems and data mining in bioinformatics, including feature reduction m