I met with Professor Verma on June 29 to discuss the problem on Gaussian Processes and my implementation of EM for GMM. I gained further clarity on the purpose of the GP problem, specifically on how kernels (and their corresponding covariance matrices) affect the shape of functions, and that the mean function is the best representation of the GP.

On the research project (which we tentatively coined Structure Preserving Metric Learning), Professor Verma suggested that I could address the zero exponent issue by perturbing the numerator and denominator with an epsilon value. However, he also thought that I should have just used an EM library function to implement my “struct” algorithm. Further clarifying the assumptions and objective of algorithm, Professor Verma provided additional guidance on how to design the “struct” loss.