Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
This lecture covers the fundamental concept of mean-square-error inference, focusing on the estimation of unaccessible parameters or random variables from related observations. It introduces the mean-square-error (MSE) criterion and discusses inference problems using the MSE criterion. The lecture explores different design criteria for inference purposes, such as mean-absolute error (MAE) and maximum a-posteriori (MAP) criteria. It delves into the inference without observations, estimation problems, and the concept of conditional mean estimator. The lecture also discusses the optimal choice for minimizing mean-square-error and the derivation of optimal estimators using completion-of-squares. It concludes with examples illustrating the application of optimal estimators in practical scenarios.