

Wasserstein gradient flows and applications to sampling in machine learning - lecture 1
De Anna Korba


Wasserstein gradient flows and applications to sampling in machine learning - lecture 2
De Anna Korba
Apparaît dans la collection : 2016 - T1 - WS4 - Inference problems theme
This talk will outline a recent set of ideas on using spatially coupled ensembles to deduce properties of the underlying non-coupled ensemble. An application is a proof of the replica symmetric formula for conditionnal entropy of Low-Density-Parity-Check codes on arbitrary binary input memoryless channels, as well as a proof of the Maxwell area construction for such systems. Applications to lossy source coding and satisfiability will be discussed time permitting.