Big data and uncertainty quantification: statistical inference and information-theoretic techniques applied to computational chemistry
mars 26 - avril 3
An incentive to use coarse-grained models is to use them for inference and control instead of the original (often intractable) model. Since coarse-grained models are always “wrong”, questions such as inference under model misspecification or goal-oriented uncertainty quantification (e.g. for control) come into play. This workshop will address such topics, with a special focus on predictive modelling, uncertainty quantification in molecular simulation and sensitivity analysis.