David Sontag's Home Page

I am a third-year graduate student in Computer Science at MIT. My research interests include machine learning and theoretical computer science, with applications in networking, computational biology, computer security, and information retrieval. I work with Tommi Jaakkola on algorithms for approximate probabilistic inference. I also work with David Karger on theoretical problems in information retrieval and networking, and with Bonnie Berger in computational biology.

I did my bachelors at UC Berkeley, in Computer Science. While there, I worked with Stuart Russell's First-Order Probabilistic Logic group.

Teaching

Publications

Probabilistic Inference:

• D. Sontag, T. Meltzer, A. Globerson, Y. Weiss, T. Jaakkola. "Tightening LP Relaxations for MAP using Message Passing". Uncertainty in Artificial Intelligence (UAI) 24, July 2008.
• D. Sontag, T. Jaakkola. "New Outer Bounds on the Marginal Polytope". Neural Information Processing Systems (NIPS) 21, Dec. 2007. Received best student paper award
• D. Sontag. "Cutting Plane Algorithms for Variational Inference in Graphical Models". Master's thesis, Massachusetts Institute of Technology, 2007.
  

  Networking:

• D. Sontag, A. Phanishayee, D. Andersen, D. Karger. "Scaling full-mesh overlay routing". Poster, NSDI, 2007.
  

  Computational Biology:

• D. Sontag, R. Singh, B. Berger. "Probabilistic Modeling of Systematic Errors in Two-Hybrid Experiments." Pacific Symposium on Biocomputing (PSB), 2007.
  

  Bayesian Logic:

• B. Milch, B. Marthi, S. Russell, D. Sontag, D. L. Ong, and A. Kolobov. "BLOG: Probabilistic Models with Unknown Objects". In Lise Getoor and Ben Taskar, eds. Statistical Relational Learning. Cambridge, MA: MIT Press, 2007.
• B. Milch, B. Marthi, S. Russell, D. Sontag, D. L. Ong, and A. Kolobov. "BLOG: Probabilistic Models with Unknown Objects". Proc. 19th International Joint Conference on Artificial Intelligence (IJCAI): 1352-1359, 2005.
• B. Milch, B. Marthi, D. Sontag, S. Russell, D. L. Ong, and A. Kolobov. "Approximate Inference for Infinite Contingent Bayesian Networks". 10th International Workshop on Artificial Intelligence and Statistics, 2005.

Past projects:

• Marginalized Kernels for Object Recognition, joint work with Brian Milch for our CS 281B and CS 294 classes (Spring 2004).
• Inference for Relational Web Page Classification, my CS 281A project report (Fall 2003).

Links:

• My father, Eduardo Sontag, and my sister, Laura Sontag.
• Lund University, in Sweden, where I studied abroad in Fall 2004.
• Upsilon Pi Epsilon, the CS honors society at Berkeley, in which I was president in Spring 2005.