1. Conceptual questions.
1 Based on the outline given in the lecture, show mathemtically that the maximum likelihood estimate (MLE) for Gaussian mean and variance parameters are given by
1 m 1 m
μˆ=m xi, σˆ2=m (xi−μˆ)2,
i=1 i=1
respectively.
Note: For this derivation, you will also need to show that these estimates for μ and σ
are maximum.
- Please compare the pros and cons of KDE as opposed to histograms, and
give at least one advantage and disadvantage to each.
- For the EM algorithm for GMM, please show how to use Bayes rule to drive
τki in closed-form expression.
2. Density estimation: Psychological experiments.
In Kanai, R., Feilden, T., Firth, C. and Rees, G., 2011. Political orientations are correlated with brain structure in young adults. Current biology, 21(8), pp.677-680., data are collected to study whether or not the two brain regions are likely to be independent of each other w.r.t. different types of political views. For this question; you can use third party histogram and KDE packages; no need to write your own implementation. The data set n90pol.csv contains information on 90 university students who participated in a psychological experiment designed to look for relationships between the size of different regions of the brain and political view. The variables amygdala and acc indicate the volume of two particular brain regions known to be involved in emotions and decision-making, the amygdala and the anterior cingulate cortex; more exactly, these are residuals from the predicted volume, after adjusting for height, sex, and similar body-type variables. The variable orientation gives the students’ locations on a five-point scale from 1 (very conservative) to 5 (very liberal). Note that in the dataset, we only have observations for orientation from 2 to 5.
1
Recall in this case, the kernel density estimator (KDE) for a density is given by 1 m 1 x i − x
where xi are two-dimensional vectors, h > 0 is the kernel bandwidth, based on the criterion we discussed in lecture. n is the number of dimensions of the KDE. For one-dimensional KDE, use a one-dimensional Gaussian kernel
2π
x1 2 x=x∈R,
2
where x1 and x2 are the two dimensions respectively
1 − (x1)2+(x2)2 K(x)= e 2 .
2π
(a) (5 points) Form the 1-dimensional histogram and KDE to estimate the distributions of amygdala and acc, respectively. For this question, you can ignore the variable orienta- tion. Decide on a suitable number of bins so you can see the shape of the distribution clearly. Set an appropriate kernel bandwidth h > 0.
- (b) (5 points) Form 2-dimensional histogram for the pairs of variables (amygdala, acc). Decide on a suitable number of bins so you can see the shape of the distribution clearly.
- (c) (10 points) Use kernel-density-estimation (KDE) to estimate the 2-dimensional density function of (amygdala, acc) (this means for this question, you can ignore the variable orientation). Set an appropriate kernel bandwidth h > 0.
Please show the two-dimensional KDE (e.g., two-dimensional heat-map, two-dimensional contour plot, etc.)
Please explain what you have observed: is the distribution unimodal or bi-modal? Are there any outliers?
Please explain based on the results, can you infer that the two variables (amygdala, acc) are likely to be independent or not?
- (d) (10 points) We will consider the variable orientation and consider conditional distri- butions. Please plot the estimated conditional distribution of amygdala conditioning on political orientation: p(amygdala|orientation = c), c = 2, . . . , 5, using KDE. Set an appropriate kernel bandwidth h > 0. Do the same for the volume of the acc: plot
2
p(x) = m i=1 hn K h ,
K(x) = √
For two-dimensional KDE, use a two-dimensional Gaussian kernel: for
1 −x2/2
e .
p(acc|orientation = c), c = 2, . . . , 5 using KDE. (Note that the conditional distribution can be understood as fitting a distribution for the data with the same orientation. Thus you should plot 8 one-dimensional distribution functions in total for this question.)
Now please explain based on the results, can you infer that the conditional distribution of amygdala and acc, respectively, are different from c = 2,…,5? This is a type of scientific question one could infer from the data: Whether or not there is a difference between brain structure and political view.
Now please also fill out the conditional sample mean for the two variables: c=2 c=3 c=4 c=5
amygdala acc
Remark: As you can see this exercise, you can extract so much more information from density estimation than simple summary statistics (e.g., the sample mean) in terms of explorable data analysis.
(e) (10 points) Again we will consider the variable orientation. We will estimate the con- ditional joint distribution of the volume of the amygdala and acc, conditioning on a function of political orientation: p(amygdala, acc|orientation = c), c = 2, . . . , 5. You will use two-dimensional KDE to achieve the goal; et an appropriate kernel band- width h > 0. Please show the two-dimensional KDE (e.g., two-dimensional heat-map, two-dimensional contour plot, etc.).
Please explain based on the results, can you infer that the conditional distribution of two variables (amygdala, acc) are different from c = 2, . . . , 5? This is a type of scientific question one could infer from the data: Whether or not there is a difference between brain structure and political view.
3. Implementing EM for MNIST dataset. [40 points]
Implement the EM algorithm for fitting a Gaussian mixture model for the MNIST hand- written digits dataset. For this question, we reduce the dataset to be only two cases, of digits “2” and “6” only. Thus, you will fit GMM with C = 2. Use the data file data.mat or data.dat. True label of the data are also provided in label.mat and label.dat.
For this problem, you must implement the EM algorithm by yourself, no third party EM implementation is allowed.
The matrix images is of size 784-by-1990, i.e., there are 1990 images in total, and each column of the matrix corresponds to one image of size 28-by-28 pixels (the image is vector- ized; the original image can be recovered by mapping the vector into a matrix).
3
First use PCA to reduce the dimensionality of the data before applying to EM. We will put all “6” and “2” digits together, to project the original data into 4-dimensional vectors.
Now implement EM algorithm for the projected data (with 4-dimensions).
- (a) (10 points) Write down detailed expression of the E-step and M-step in the EM algo- rithm. (hint: when computing τki, you can drop the (2π)n/2 factor from the numerator and denominator expression, since it will be canceled out; this can help avoid some numerical issues in computation).
Note: It is not sufficient to use N (…) to represent the multivariate Gaussian, you must fully write out any p.d.f. expressions.
- (b) (15 points) Implement EM algorithm yourself. Use the following initialization initialization for mean: random Gaussian vector with zero mean
initialization for covariance: generate two Gaussian random matrix of size n-by- n: S1 and S2, and initialize the covariance matrix for the two components are Σ1 = S1S1T + In, and Σ2 = S2S2T + In, where In is an identity matrix of size n-by-n.
Plot the log-likelihood function vs the number of iterations to show your algorithm is converging.
- (c) (5 points) Report the fitted GMM model when EM terminates. For the mean of each component, map these back to the original space and reformat the vectors to make them into 28-by-28 matrices and show images. Ideally, you should be able to see these means correspond to “average” images. You can report the two 4-by-4 covariance matrices by visualizing their intensities (e.g., using a gray scaled image or heat map).
- (d) (10 points) Use τki to infer the labels of the images, and compare with the true labels. Report the mis-classification rate for digits “2” and “6” respectively. Perform K-means clustering with K = 2 (you may call a package or use code from previous assignments). Find the mis-classification rate for digits “2” and “6” respectively, and compare with GMM. Which model achieves better performance overall?
4
