Project 0: Experiments on Deep Learning with Conv. Neural Network

 1. Objectives.
    In the past 3 years, deep learning has become popular and has been used widely for pattern classification tasks. This project is designed to provide
 you first hand experience with training and testing a typical convolutional neural network structure on a task of classifying 10 object categories.
    But, don't panic, you don't have to really understand the algorithm and learning method, as standard code and data are provided to you. All you 
need to do is to follow the steps and run some experiments, then report results.   

 2. Datasets.
  We will use the CFAR-10 dataset for its reasonable size. The data includes 60,000 images for 10 object categories with 6,000 each category, 
and each image has 32x32 pixels. 50,000 for training and 10,000 for testing. A website for this dataset is here. 

 3. Instructions and code. 
  Download the Instructions (pdf), dataset (zip),  the code (downloaded from training and testing)
  and code in matlab(zip).

  TA will provide a tutorial session (location and room to be announced).