The goal of this assignment is to give you a concrete understanding of the Tenenbaum (1999) work by implementing and experimenting with a small scale version of the model applied to learning concepts in two-dimensional spaces.  The further goal is to get hands-on experience representing and manipulating probability distributions and using Bayes' rule.

Consider a two-dimensional input space with features in the range [-10, 10]. We will consider only square concepts, and concepts centered on the origin (0,0).  We will also consider only a discrete set of concepts, H = {hi, i=1...10}, where hi is the concept with lower left corner (-i, -i) and upper right corner (+i, +i), i.e., a square with the length of each side being 2i.

You will have to define a discrete prior distribution over the 10 hypotheses, and you will have to do prediction by marginalizing over the hypothesis space.  Use Tenenbaum's expected-size prior.  Because the expected size prior is defined over a continuous distribution, you will need to compute the value of the prior for each of the 10 hypotheses, and renormalize the resulting probabilities so that the prior distribution sums to 1.  (You don't actually need to do this renormalization, because the normalization factor cancels out when you do the Bayesian calculations, but go ahead and do it anyhow, just to have a clean representation of the priors.)

Make a bar graph of the prior distribution for sigma1=sigma2 = 2.  Make a graph of the prior distribution for sigma1 = sigma2 = 5.

Compute posteriors over H, P(H|X) for X = {(1.5, .5)} and sigma = 10. You will get one probability for each possible hypothesis. Display your result either as a bar graph or a list of probabilities.

Using the results of Task 2, compute generalization predictions, P(y|X), over the whole input space for X = {(1.5, .5)} and sigma = 10. The input space should span the region from (-10,-10) to (+10,+10). Display your result as a contour map in 2D space where the coloring of the contour map represents the probability that an input at that point in the space will be a member of the concept. (If the probabilities are becoming very small, you can display log(probability) to flatten the distribution.)

Repeat Task 3 for sigma = 1.

Compute generalization predictions, P(y|X), over the whole input space for sigma = 10 and three different sets of input examples: X = {(2.2, -.2)}, X = {(2.2, -.2), (.5, .5)}, and X = {(2.2, -.2), (.5, .5), (3.5, 1)}.

Do some other interesting experiment with the model.  One possibility would be to extend the model to accommodate negative as well as positive examples.  Another possibility would be to compare generalization surfaces with and without the size principle, and with an uninformative prior (here, uniform would work) compared to the expected-size prior.

I would like hardcopies of your work.   It's easier for you than putting together a single word / pdf document.  And it's easier for me than keeping tabs on electronic documents.  It would be good for you to hand in the code in case there's a problem with your results.