Likelihood Based

Likelihood-based methods are central to statistical inference, aiming to estimate parameters or make predictions by maximizing the probability of observed data given a model. Current research focuses on addressing computational challenges in high-dimensional or complex models, employing techniques like neural networks to approximate intractable likelihoods, and using alternative approaches such as scoring rules or optimal transport to bypass direct likelihood calculations. These advancements are improving the efficiency and applicability of likelihood-based inference across diverse fields, from medical image analysis and astrophysics to robust control of bioprocesses and improved machine learning model generalization.