Wednesday, June 19, 2019 - 4:15pm - 5:05pm
Moritz Hardt (University of California, Berkeley)
We clarify what fairness guarantees we can and cannot expect to follow from unconstrained machine learning. Specifically, we characterize when unconstrained learning on its own implies group calibration, that is, the outcome variable is conditionally independent of group membership given the score. We show that under reasonable conditions, the deviation from satisfying group calibration is upper bounded by the excess risk of the learned score relative to the Bayes optimal score function. A lower bound confirms the optimality of our upper bound.
Tuesday, June 1, 2010 - 3:30pm - 4:15pm
Scott Kelly (University of North Carolina)
Models for aquatic locomotion generally seek to balance fidelity and scope with analytical or computational tractability. When the goal in model development is a platform for model-based feedback control design, analytical structure is essential to provide a point of access for most current design techniques, but some fidelity may be sacrificed as long as the scope of the model encompasses the range of situations under which control will be applied.
Friday, May 7, 2010 - 1:25pm - 2:25pm
Genetha Gray (Sandia National Laboratories)
Despite their obvious advantages, computer simulations also introduce many challenges. Uncertainties must be identified and quantified in order to guarantee some level of predictive of a computational model. Calibration techniques can be applied to both improve the model and to curtail the loss of information caused by using simulations in place of the actual system.
Tuesday, December 17, 2013 - 10:25am - 10:55am
Paul Patrone (University of Minnesota, Twin Cities)
In recent years, the composites community has increasingly used molecular dynamics to simulate and explore material properties such as glass-transition temperature and yield strain. In virtually all such simulations, a key challenge is to select one or more input structures that represent the real polymer matrix at the nanoscale. Often an appropriate choice of inputs is not known a priori, which can lead to a large uncertainty in the simulated composite properties.
Subscribe to RSS - Calibration