On financial markets one never observes the same data twice; market configurations are subject to change across time. This poses some specific challenges to inference, prediction, and optimal control in financial contexts. Classically, strong model assumptions are needed, while current research aims at methods which are robust with respect to model misspecification. This issue lies at the heart of the envisaged workshops, and the program of the workshops will reflect recent developments in this direction. The last decade saw a rise of robust methods in probability and finance resulting in new numerical and theoretical challenges. Interestingly, these challenges bring together methodologies from PDEs, probability, stochastic analysis, and control theory. Mathematically speaking, robustness typically translates into nonlinearity showing up as a defining feature. Examples in this direction are nonlinear expectations, nonlinear PDEs, and H-infinity optimal stochastic control. Finance has a long tradition of fruitful interactions between these areas. Numerical results often build the first step for subsequent theoretical analysis (and vice versa), thus fitting specifically into ICERM's orientation towards computational and experimental research.