# CS 179: Introduction to Graphical Models Homework 7

Last Modified: May 24, 2022

CS 179: Introduction to Graphical Models: Spring 2022
Homework 7
Due Date: Friday, June 3, 2022
The submission for this homework should be a single PDF file containing all of the relevant code, figures, and any
text explaining your results. When coding your answers, try to write functions to encapsulate and reuse code,
instead of copying and pasting the same code multiple times. This will not only reduce your programming efforts,
but also make it easier for us to understand and give credit for your work. Show and explain the reasoning
behind your work!
In this homework, we will run a simple variational auto-encoder (VAE) model and explore its resulting
representation.
To simplify the effort for this homework, a template containing much of the required code for the assignment
is provided in a Jupyter ipynb file.
Part 1: Build the VAE & load data (30 points)
First, download the template Jupyter notebook and look over the provided code. You can run the code locally or
on Google Colab, as you prefer. The VAE is defined by two quantities. First, an encoder defines the variational
distribution q(Z|X = x), which we express as a Gaussian distribution,
Z ∼ N ￾ Z;µ(x),ν(x)I
 (µ,logν) = W2 α(W1 x )
where α(·) is a ReLU activation function and is the usual matrix-vector dot product. In other words, (µ,logν)
are expressed by a two-layer neural network with ReLU activation on the hidden layer, and linear activation on the
output.
The decoder defines p(X|Z = z), which we express as a Gaussian distribution,
X ∼ N ￾ X; ¯µ(z),ω
 ¯µ(z) = σ( V2 α( V1
z ) )
where σ(·) is the logistic function, so that X is modeled as a two-layer neural network transformation of z, plus a
fixed amount of Gaussian noise on the pixel values.
The loss is given by the divergence between ˆp(X)q(Z|X) and p(Z)p(X|Z), where we assume p(Z) is a basic
unit Gaussian, which we estimate by sampling from q(Z|X). Given samples {(x
(i)
, z
(i)
)}, our estimated loss is
1
m
k
¯µ(z
(i)
) − x
(i)
k 2 +
1
2
detν(x
(i)
) + k µ(x
(i)
)k
2 − log detν(x
(i)
) − 1

Data Our data consist of a small sample of hand images (mine, but inspired by Josh Tenenbaum’s IsoMap
experiment), located at
https://sli.ics.uci.edu/extras/cs179/data/frames.txt
Load the data, then shift and scale the values to be between zero and one:
1 data -= data.min(1,keepdims=True)
2 data /= data.max(1,keepdims=True)
3 data = torch.tensor(data).float()
Part 2: Train the model (30 points)
You have also been provided with a function train which computes the gradient of a mini-batch of data and
updates the parameter values. Use this function to train your model:
1 vae = VAE()
2 optim = Adam(vae.parameters(), lr=0.0001)
3 train(vae, data, optim, batch = 16, epochs=500)
Homework 7 UC Irvine 1/ 2
CS 179: Introduction to Graphical Models Spring 2022
This may be a bit slow. The train function also calls another provided function, plot_scatter , which plots
a scatter of images (preventing any overlapping images), so you can visualize the two-dimensional latent space Z
which is being used to capture the variability in the images.
Part 3: Visualize reconstructions (30 points)
Note: if you are short on time, or want to compare your results, you can obtain my trained VAE at
https://sli.ics.uci.edu/extras/cs179/data/vae.pkl
and load using pickle ,
1 with open('vae.pkl','rb') as fh: vae = pickle.load(fh)
However, if you have trained your own in part (2), please use yours for this question as well.
(a) Select 6 random images from the data set. Encode and then decode each image, and show ( imshow ) the
original image and reconstructed image, ¯µ(µ( x ) ). (Note: these will not look that great; the reconstruction
network may be too simple to do a good job, or perhaps we just don’t have enough data in this small data
set. But they should be recognizable.)
(b) Now select 10 points in a linear path across the distribution (say, from z = (−3,0) to z = (3,0). For each
latent location, decode ¯µ(z). Interpret the resulting sequence of images.
Part 4: Work on your projects! (10 points)
Good luck with your projects, and your finals for any other classes!
Homework 7 UC Irvine 2/ 2

© 2024 www.7daixie.com