Recent progress in machine learning provides many potentially effective tools to learn estimates or make predictions from datasets of ever-increasing sizes. Can we trust such tools in clinical and highly-sensitive systems? If a learning algorithm predicts an effect of a new policy to be positive, what guarantees do we have concerning the accuracy of this prediction? The talk introduces new statistical ideas to ensure that the learned estimates satisfy some fundamental properties: especially causality and robustness. The talk will discuss potential connections and departures between causality and robustness.
Jelena Bradic is an Associate Professor at the UC San Diego, where she holds a joint appointment in the Department of Mathematics and Halicioglu Data Science Institute.
Prof. Bradic holds a Ph.D. degree in Statistics from Princeton University (2011) associated with the Operations Research and Financial Engineering Department . She has had the pleasure to study under the guidance of Prof Jianqing Fan. Her undergraduate and masters degree are in Mathematics from Belgrade University, Serbia (2004, 2007).
Prof. Bradic's interests are in causal inference, machine learning, robust statistics as well as missing data problems. Her application areas include observational and interventional data, treatment effects, as well as public health and policy learning. She strives to understand and develop new robust learning methods and algorithms with provable guarantees of stability, robustness to data corruption and data generating mechanism.