Machine Learning Speaking Skills Talk

Physicists frequently encounter the problem of estimating important cosmological constants such as the Hubble constant or the amount of dark matter in the Universe. Usually this task is cast in the Bayesian setting where we assume a prior on these quantities and the likelihood is given by a physical model. The task at hand is to estimate the posterior given some astronomical observations. Cosmological simulators can be used to estimate the likelihood for a given value of these constants. However, these simulations are very expensive and conventional methods for posterior estimation breakdown. In this work we study active posterior estimation when the likelihood is expensive to evaluate. Existing techniques for posterior estimation are based on generating samples representative of the posterior. Such methods do not consider efficiency in terms of…


Link to Full Article: Machine Learning Speaking Skills Talk