Conveners
ODSL Highlights
- Allen C. Caldwell (MPP)
A common thread of the ORIGINS is the presence of ever increasing data volumes both from instruments and simulation that simultaneously promise wealth of scientific insight. This data-intensive era, which often can only be managed in large scientific collaborations, coincides with a particularly interesting, but challenging phase in computing: While Machine Learning, particularly Deep...
In this talk we present how to combine independently trained neural networks to jointly solve novel tasks through Bayesian reasoning. Deep generative networks serve as prior distributions on complex systems and regression/classification networks are used to check whether certain features are present. Bayes Theorem allows us to then solve the inverse problem in terms of the latent variables of...