• Deep Learning Concepts for Evolutionary Art

      Tanjil, Fazle; Department of Computer Science
      A deep convolutional neural network (CNN) trained on millions of images forms a very high-level abstract overview of any given target image. Our primary goal is to use this high-level content information of a given target image to guide the automatic evolution of images. We use genetic programming (GP) to evolve procedural textures. We incorporate a pre-trained deep CNN model into the fitness. We are not performing any training, but rather, we pass a target image through the pre-trained deep CNN and use its the high-level representation as the fitness guide for evolved images. We develop a preprocessing strategy called Mean Minimum Matrix Strategy (MMMS) which reduces the dimensions and identifies the most relevant high-level activation maps. The technique using reduced activation matrices for a fitness shows promising results. GP is able to guide the evolution of textures such that they have shared characteristics with the target image. We also experiment with the fully connected “classifier” layers of the deep CNN. The evolved images are able to achieve high confidence scores from the deep CNN module for some tested target images. Finally, we implement our own shallow convolutional neural network with a fixed set of filters. Experiments show that the basic CNN had limited effectiveness, likely due to the lack of training. In conclusion, the research shows the potential for using deep learning concepts in evolutionary art. As deep CNN models become better understood, they will be able to be used more effectively for evolutionary art.
    • Using Deep Learning for Predicting Stock Trends

      Fazeli, Arvand; Department of Computer Science
      Deep learning has shown great promise in solving complicated problems in recent years. One applicable area is finance. In this study, deep learning will be used to test the predictability of stock trends. Stock markets are known to be volatile, prices fluctuate, and there are many complicated financial indicators involved. While the opinion of researchers differ about the predictability of stocks, it has been shown by previous empirical studies that some aspects of stock markets can be predictable to some extent. Various data including news or financial indicators can be used to predict stock prices. In this study, the focus will be on using past stock prices and using technical indicators to increase the performance of the results. The goal of this study is to measure the accuracy of predictions and evaluate the results. Historical data is gathered for Apple, Microsoft, Google and Intel stocks. A prediction model is created by using past data and technical indicators were used as features in the model. The experiments were performed by using long short-term memory networks. Different approaches and techniques were tested to boost the performance of the results. To prove the usability of the final model in the real world and measure the profitability of results backtesting was performed. The final results show that while it is not possible to predict the exact price of a stock in the future to gain profitable results, deep learning can be used to predict the trend of stock markets to generate buy and sell signals.