Social Decision and Analytics Laboratory

Abstract:  The common use of computational models, in combination with physical observations, has expanded our understanding and ability to anticipate behaviors in a variety of physical systems. With relevant physical observations, it is possible to calibrate a computational model, and even estimate systematic discrepancies between the model reality. Estimating and quantifying the uncertainty in this model discrepancy can lead to reliable prediction uncertainties - so long as this prediction is “similar” to the available physical observations. Exactly how to define \"similar\" has proven difficult in many applications. Clearly it depends on how well the computational model captures the relevant physics in the system, as well as the portability of the model discrepancy in moving from the available physical data to the prediction. This talk will discuss these concepts using computational models ranging from simple to complex.