Understanding the flow of information in Deep Neural Networks (DNNs) is a challenging problem that has gain increasing attention over the last few years. While several methods have been proposed to explain network predictions, only a few attempts to analyze them from a theoretical perspective have been made in the past. In this work, we analyze various state-of-the-art attribution methods and p...

this article is a comparative study of estimation power of artificial neural networks and autoregressive time series models in inflation forecasting. using 37 years iran’s inflation data, neural networks performs better on average for short horizons than autoregressive models. this study shows usefulness of early stopping technique in learning stage of neural networks for estimating time series...

Picasso is a free open-source (Eclipse Public License) web application written in Python for rendering standard visualizations useful for analyzing convolutional neural networks. Picasso ships with occlusion maps and saliency maps, two visualizations which help reveal issues that evaluation metrics like loss and accuracy might hide: for example, learning a proxy classification task. Picasso wor...

Email has become an essential communication tool in modern life, creating the need to manage the huge information generated. Email classification is a desirable feature in an email client to manage the email messages and categorize them into semantic groups. Statistical artificial intelligence and machine learning is a typical approach to solve such problem, driven by the success of such method...

We have developed a deep learning network for classification of different flowers. For this, we have used Visual Geometry Group’s 102 category flower data-set having 8189 images of 102 categories from Oxford University. The method is basically divided in two parts i.e. Image segmentation and classification. We have compared two different Convolutional Neural Network architectures GoogLeNet and ...

Using deep learning to improve the capabilities of high-resolution satellite images has emerged recently as an important topic in automatic classification. Deep networks track hierarchical high-level features to identify objects; however, enhancing the classification accuracy from low-level features is often disregarded. We therefore proposed a two-stream deep-learning neural network strategy, ...

—Deep neural networks are state-of-the-art models for understanding the content of images, video and raw input data. However, implementing a deep neural network in embedded systems is a challenging task, because a typical deep neural network, such as a Deep Belief Network using 128×128 images as input, could exhaust Giga bytes of memory and result in bandwidth and computing bottleneck. To addre...

Advances in supervised learning with deep neural networks have enabled robust classification in many real world domains. An interesting question is if such advances can also be leveraged effectively for computational creativity. One insight is that because evolutionary algorithms are free from strict requirements of mathematical smoothness, they can exploit powerful deep learning representation...

We introduce FFTNet, a deep learning approach synthesizing audio waveforms. Our approach builds on the recent WaveNet project, which showed that it was possible to synthesize a natural sounding audio waveform directly from a deep convolutional neural network. FFTNet offers two improvements over WaveNet. First it is substantially faster, allowing for real-time synthesis of audio waveforms. Secon...