Assignment作业 | 机器学习 | 算法代写 | 代写python- 这是一个涉及神经网络的算法题目代写
assignment 4: Convolutional Neural Networks
CS489/698 Winter 2018
Be sure to include your name and student number with your assignment.
- [50 pts]For this question, you will experiment with fully connected neural networks and convolutional neural networks, using the Keras open source package. Keras is one of the simplest deep learning package that serves as a wrapper on top of TensorFlow, CNTK and Theano. Preliminary steps: – Download and install Keras fromhttps://keras.io/. A CPU installation is sufficient for this as- signment. – Click on Getting Started and read the Guide to the Sequential Model. – Download the filecifar10cnn.pyfrom the example folder https://github.com/keras-team/keras/tree/master/examples.
Answer the following questions by modifying the code incifar10cnn.py.
(a) Compare the accuracy of the convolutional neural Network in the filecifar10cnn.pyon the cifar
dataset to the accuracy of simple dense neural networks with 0, 1, 2, 3 and 4 hidden layers of 512 rectified
linear units each. Modify the code incifar10cnn.pyto obtain simple dense neural networks with 0,
1, 2, 3 and 4 hidden layers of 512 rectified linear units (with a dropout rate of 0.5). Produce a graph that
contains 6 curves (one for the convolutional neural net and one for each dense neural net of 0-4 hidden
layers). The y-axis is the test (validation) accuracy and the x-axis is the number of epochs (# of passes
through the training set). Produce curves for the first 10 epochs. Although 10 epochs is not sufficient to
reach convergence, it is sufficient to see the trend. Explain the results (i.e., why some models perform
better or worse than other models). No need to submit your code since the modifications are simple.
(b) Compare the accuracy achieved by rectified linear units and sigmoid units in the convolutional neural
network incifar10cnn.py. Modify the code incifar10cnn.pyto use sigmoid units. Produce a
graph that contains 2 curves (one for rectified linear units and another one for sigmoid units). The y-axis
is the test (validation) accuracy and the x-axis is the number of epochs (# of passes through the training
set). Produce curves for the first 10 epochs. Although 10 epochs is not sufficient to reach convergence,
it is sufficient to see the trend. Explain the results (i.e., why did one model perform better than the other
model). No need to submit your code since the modifications are simple.
(c) Compare the accuracy achieved with and without drop out as well as with and without data augmentation
in the convolutional neural network incifar10cnn.py. Modify the code incifar10cnn.pyto
turn on and off dropout as well as data augmentation. Produce two graphs (one for training accuracy and
the other one for test accuracy) that each contain 4 curves (with and without dropout as well as with and
1
without data augmentation). The y-axis is the accuracy (i.e., train or test/validation accuracy) and the x-
axis is the number of epochs (# of passes through the training set). Produce curves for as many epochs as
you can up to 100 epochs. Explain the results (i.e., why did some models perform better or worse than
other models and are the results consistent with the theory). No marks will be deducted for doing less than
100 epochs, however make sure to explain what you expect to see in the curves as the number of epochs
reaches 100. No need to submit your code since the modifications are simple.
- [50 pts]In object recognition, translating an image by a few pixels in some direction should not affect the category recognized. Suppose that we consider images with an object in the foreground on top of a uniform background. Suppose also that the objects of interest are always at least 10 pixels away from the borders of the image. Are the following neural networks invariant to translations of at most 10 pixels in some direction? Here the translation is applied only to the foreground object while keeping the background fixed. If your answer is yes, show that the neural network will necessarily produce the same output for two images where the foreground object is translated by at most 10 pixels. If your answer is no, provide a counter example by describing a situation where the output of the neural network is different for two images where the foreground object is translated by at most 10 pixels.
(a)[25 pts]Neural network with one hidden layer consisting of convolutions (5x5 patches with a stride of 1
in each direction) and a softmax output layer.
(b)[25 pts]Neural network with two hidden layers consisting of convolutions (5x5 patches with a stride of 1
in each direction) followed by max pooling (4x4 patches with a stride of 4 in each direction) and a softmax
output layer.