Abstract: We begin by examining recent advancements in statistical learning when dealing with contaminated and heavy-tailed high-dimensional data. We prove the existence of estimators with optimal statistical performance and discuss the phenomenon of overfitting. Moving forward, we explore statistical aspects of high-dimensional optimal transport and present methods for overcoming the curse of dimensionality.
Zoom-Link:
https://univienna.zoom.us/j/63667270827?pwd=Rkg2ODdqeTBtQTMwNWFYdVo3by9KZz09