Abstract:
In this dissertation, application of soft computing techniques in speech recognition has been explored. At present the prevailing technology for speech recognition is predominantly Hidden Markov Model based. a statistical framework that supports both acoustic and temporal modeling. Despite their state-of-the-art performance, HMMs make a number of sub optimal modeling assumptions \ that limit their potential effectiveness. Neural networks avoid many of these assumptions, while they can also learn complex functions, generalize effectively, tolerate noise, and support parallelism.
Neural networks can be used for situations where speech feature vectors are non-linearly distorted, such as in noisy reverberant speech or telephone speech. By using a neural network, the adaptation process requires a small amount of training data. First, a neural network is applied to the computation of an inverse distortion function. This type of network requires simultaneously recorded input and target pairs for training. Traditionally, neural networks are trained to minimize the mean squared error between the network output and the corresponding target value. However, minimizing the mean squared error does not guarantee maximum recognition accuracy. Therefore, a new objective function for the neural network is proposed, which makes use of fuzzy rules.
Speech recognition throws up a myriad of problems, these problems are generally of pattern recognition, approximation and optimization. These problems have imprecise and distorted data. As a result conventional techniques are grossly inadequate in this domain. Fuzzy logic is the most appropriate in these circumstances. Optimization is necessary while designing Neural Network and handling the input features. This dissertation has explored the application of this paradigms under various conditions
