Erik Zawadzki

Research

I am advised by the following unordered pair: {Geoff Gordon, André Platzer}. Currents interests are:

complementarity problems
convex optimization and analysis
planning and MDPs
learning theory and machine learning
probabilistic inference

I am currently working on methods for incorporating function approximation into solution methods for a broad class of linear complementarity problems. These complementarity problems are strongly connected to the Karush-Kuhn-Tucker conditions for linear and quadratic programs, but are more general. Applications include approximating Markov decision processes and approximating support vector machines. Approximate solution methods include projected-gradient descent, proximal methods, and interior point methods.

I am a member of the Logical Systems Lab and the SELECT Lab. I have also been helping a CMU/JHU APL team out with verifying the FAA's new Airborne Collision Avoidance System (ACAS X), which is based on an MDP.

Education

Ph.D. in Computer Science from Carnegie Mellon University (2008 - )
M.Sc. in Computer Science from the University of British Columbia; advised by Kevin Leyton-Brown (2007 - 2008)
EPIC Intern, IBM Toronto (2004 - 2005)
B.Sc. Honours in Computer Science from the University of British Columbia (2001 - 2006)

Papers

"Nonparametric Scoring Rules." Erik Zawadzki, Sébastien Lahaie. AAAI, 2015. (Paper)
"A Formally Verified Hybrid System for the Next-Generation Airborne Collision Avoidance System." Jean-Baptiste Jeannin, Khalil Ghorbal, Yanni Kouskoulas, Ryan Gardner, Aurora Schmidt, Erik Zawadzki, and André Platzer. TACAS, 2015. (Paper)
"Hybrid theorem proving of aerospace systems: Applications and challenges." Khalil Ghorbal, Jean-Baptiste Jeannin, Erik Zawadzki, André Platzer, Geoffrey Gordon, and Peter Capell. Journal of Aerospace Information Systems, 2014. (Paper)
"A Projection Algorithm for Strictly Monotone Linear Complementarity Problems." Erik Zawadzki, Geoffrey Gordon, André Platzer. NIPS OPT, December 2013. (Paper)
"A Generalization of SAT and #SAT for Robust Policy Evaluation." Erik Zawadzki, André Platzer, Geoffrey Gordon. IJCAI, August 2013. (Paper)
"Memory-Efficient GroupBy-Aggregate with Compressed Buffer Trees." Hrishikesh Amur, Wolf Richter, David Andersen, Michael Kaminksy, Karsten Schwan, Athula Balachandran, Erik Zawadzki. ACM Symposium on Cloud Computing, October 2013 (Paper)
"An Instantiation-Based Theorem Prover for First-Order Programming." Erik Zawadzki, Geoffrey Gordon, André Platzer. AISTATS, April 2011. (Paper)
"Search Tree Restructuring." Erik Zawadzki, Tuomas Sandholm. Technical Report, May 2010. (Paper)
"Empirically Evaluating Multiagent Learning Algorithms." Erik Zawadzki, Ascher Lipson, Kevin Leyton-Brown. Working Paper, November 2014. (arXiv)
"Empirically testing decision making in TAC SCM.", Erik Zawadzki, Kevin Leyton-Brown. AAAI-07 Workshop on Trading Agent Design and Analysis (TADA-07), Vancouver, 2007. (Paper)
"Multiagent learning and empirical methods." Erik Zawadzki. Master's Thesis, 2008. (Thesis)