We develop novel methods to make Bayesian inference more efficient, scalable, and practical. This includes work on variational methods, Monte Carlo algorithms, and techniques for handling complex ...