International Conference on Research Trends in Computer Technologies (ICRTCT - 2013) Proceedings published in International Journal of Computer Applications® (IJCA) (0975 – 8887) 5 Co-relation based approach for Pattern Recognition using ANN and its Fault Tolerance Analysis Farhana Kausar V Vijayalakshmi AtriaInstitute of Technology Atria Institute of Technology, ASKB Campus, Anadanagar, Bangalore. ASKB Campus, Anadanagar, Bangalore ABSTRACT The best pattern recognizers in most instances are human, yet we do not understand how human recognize patterns. The pattern recognition is critical in the human decision task, the more relevant the pattern at your disposal, the better your decision will be. More recently, neural network techniques in pattern recognition have been receiving increasing attention. The design of a recognition system requires careful attention to the following issues: definition of pattern classes, sensing environment, pattern representation, feature extraction and selection, learning, selection of training and test samples. A review of fault tolerance in neural networks is presented. Work relating to the various issues of reliability, complexity and capacity are considered, as well as covering both empirical results and a general treatment of theoretical work. It is shown that in the majority of the work, few sound theoretical methods to be applied and that conventional fault tolerance techniques cannot straightforwardly be transferred to neural networks. It is often concluded that all the neural networks are often cited as being fault tolerant. To support this work a fundamental prerequisite is described which can act as base for research into the fault tolerance of neural networks. (The performance analysis of fault tolerance is done using uniform and Gaussian distribution.) KEYWORDS Feedforward Neural Networks,, Fault Tolerance, Fault Model, Graceful Degradation, Pattern Recognition Introduction An Artificial Neural Network (ANN) is an information processing paradigm that is inspired by the way biological nervous systems, such as the brain, process information. The key element of this paradigm is the novel structure of the information processing system. It is composed of a large number of highly interconnected processing elements (neurons) working in unison to solve specific problems. ANNs, like people, learn by example. An ANN is configured for a specific application, such as pattern recognition or data classification, through a learning process. Learning in biological systems involves adjustments to the synaptic connections that exist between the neurons. This is true of ANNs as well. In the case of biological neural network, tolerance to loss of neurons has high priority, since agraceful degradation of performance is very important for survival of the organism. Faulttolerance measures the capacity of neural network to perform the desired task under given faultcondition. It also maintains their computing ability when a part of the network is damaged orremoved. Fault tolerance is one of the key performances of artificial neural networks (ANN‟s) and is often viewed as an inherent feature of ANN‟s. But without precise designing it is not able to guarantee the degree of fault tolerance. A very precise recognition of learned patterns even if there is a variation in the applied test patterns. Analysis of fault tolerance properties in ANN with feed-forward architecture using bit map images. The analysis of fault tolerance with respect to Uniform and Gaussian/Normal distribution with respect to hidden layer, output layer, both hidden and output layer is done graphically. Fault tolerances of ANN have been studied in [1][15]. Fault tolerance of ANN may be characterized/categorized on the following aspects: (i) Weight error: Weight stuck at zero/max/min (ii) Neuron error: Node stuck at zero/max/min (iii) Input pattern errors: injecting noise during the training phase. To design an ANN, which could recognize the learned patterns even if there is a variations in applied test patterns.The analysis of fault tolerant property using designed ANN. Effects of faults at different position. 1. Hidden layer 2. Output layer 3. With processing elements at hidden layer Fault tolerance is one of the key performances of artificial neural networks (ANN‟s) and is often viewed as an inherent feature of ANN‟s. But without precise designing it is not able to guarantee the degree of fault tolerance. A very precise recognition of learned patterns even if there is a variation in the applied test patterns. Analysis of fault tolerance properties in ANN with feed-forward architecture using bit map images. The analysis of fault tolerance with respect to Uniform and Gaussian/Normal distribution with respect to hidden layer, output layer, both hidden and output layer is done graphically. Related Work With the widespread usage of the chip-based device of the ANN as controller, it has become imperative to study the behavior of these circuits under various faults, i.e., study of their fault tolerance behavior must be undertaken. The available literature on the fault-tolerance behavior offeedforward ANNs may be summarized as: 1. Neural networks are not intrinsically fault tolerant and their fault tolerance has to be improved by employing extra mechanism.[1] 2. Artificial neural networks possess characteristics that make them a useful tool for pattern recognition and classification. Important features such as