Bio

I am a senior at Columbia University studying both computer science and mathematics. My interests within these fields falls mainly into the realm of computer science theory. In particular, I am interested in cryptography, communication complexity, and formal verification.

Current Research

I am currently working with Professor Suman Jana in the field of abstract interpretation to create a general abstract domain refinement procedure based on the notion of interval refinement. Such a generalized splitting scheme is useful in the realm of dynamic program analysis, as many properties of programs can be modeled as geometric constraints on inputs and outputs. I am also currently working with Professor Allison Bishop to create a cryptographic bilinear map based on the learning with errors problem.

Past Research

From September 2017 until August 2018, I worked with Professor Jana and graduate student Shiqi Wang on a project aimed at using interval analsis to verify security properties of neural networks against adversarial examples, small input perturbations that lead to misclassifications. Results from have project have ended up in both USENIX Security 2018 and NIPS 2018. I have also worked with Dr. Kyle Chard to develop a research repository with which scientists entering a new field can quickly find literature relevant to their research.

Publications

  1. S. Wang, K. Pei, J. Whitehouse, J. Yang, S. Jana. Efficient Formal Safety Analysis of Neural Networks. Conference on Neural Information Processing Systems (NIPS), 2018.

  2. S. Wang, K. Pei, J. Whitehouse, J. Yang, S. Jana. Formal Security Analysis of Neural Networks Using Symbolic Intervals. 27th USENIX Security Symposium, 2018.