'What one fool could understand, another can.'
                                                                                                                                                  -- R.P. Feynman

The course will give participants an opportunity to implement statistical models. We will cover numerical optimization techniques including gradient descent, newton's method and quadratic programming solvers to fit linear and logistic regression, discriminant analysis, support vector machines and neural networks. The second part of the course will focus on advanced topics for computing posterior distributions and motivate their appeal in Bayesian inference. We will survey importance and rejection sampling, Metropolis, Gibbs sampling, and Sequential Monte Carlo. Students will be exposed to convex duality, constrained optimization, bias/variance decompositions, entropy, mutual information, KL divergence, maximum likelihood/maximum a posteriori estimation, Fisher scoring, Laplace approximation, Markov chains and saddle point methods, each of which will be reemphasized from a computational perspective.