Despite the remarkable success of deep learning in pattern recognition, deep network models face the problem of training a large number of parameters. In this paper, we propose and evaluate a novel multi-path wavelet neural network architecture for image classification with far less number of trainable parameters. The model architecture consists of a multi-path layout with several levels of wavelet decompositions performed in parallel followed by fully connected layers. These decomposition operations comprise wavelet neurons with learnable parameters, which are updated during the training phase using the back-propagation algorithm. We evaluate the performance of the introduced network using common image datasets without data augmentation except for SVHN and compare the results with influential deep learning models. Our findings support the possibility of reducing the number of parameters significantly in deep neural networks without compromising its accuracy.

No abstract

No abstract

No abstract

The cost of unserved energy is an important parameter in long term generation planning. Indices produced based on this value are used in a wide range of management decisions throughout the utility. The value of cost of unserved energy currently used by Ceylon Electricity Board (CEB) is based on studies conducted over a decade ago. It is important to update such estimates. The paper presents an analysis of data on electricity interruptions for different industry categories based on their contribution to the economy and losses caused by electricity interruptions is presented. This data was then used to calculate cost of energy unserved in the selected industry categories. Data was gathered through a survey based study conducted on a stratified sample to cover the full spectrum of industries. The work included design of the questionnaire, selection of the sample, conduct of the survey, analysis and conclusions. The study developed a satisfactory methodology for calculating the cost of unserved energy.