In 2012, Eric began researching with Zico Kolter on the problem of molecular energy optimization, developing specialized kernels for geometrically structured data, which culminated in publishing a paper at AAAI in 2015. Over the same period of time, Eric worked at CERT on security problems (migrating security rules to newer C standards and analyzing program vulnerabilities). Eric is now pursuing a PhD in Machine Learning and has become a leader in the field of provable defenses against adversarial examples, providing the first deep networks with certified guarantees under adversarial perturbations. He is currently interning at Bosch to bring advancements in adversarial examples into the automotive industry and robustify real sensor systems, both visual systems (i.e. camera based) and physical systems (i.e. fuel injection controllers). He continues to work at the intersection of robustness, deep learning, and convex optimization, bringing principled approaches to the often empirical nature of adversarial examples.
Ph.D. , Computer ScienceClass of 2020