Brenda Ng Computer Science Ph.D. Affiliation: Harvard AI Group Degree advisor: Avi Pfeffer Contact me by email

I studied under Prof. Avi Pfeffer while I was a graduate student in the Department of Engineering and Applied Sciences at Harvard University. My dissertation research specialized in probabilistic reasoning using sequential Monte Carlo methods. In particular, my research was on factored inference of complex dynamic systems, whereby I explored ways of combining parametric and non-parametric ways of density estimation within the framework of discrete- and continuous-time inference.

Research Interests

My research interest is in the design and real-time application of approximate probabilistic inference algorithms that can efficiently monitor and diagnose complex dynamic systems. In particular, I am interested in:

• efficient, any-real-time approximate reasoning of large hybrid-state systems whose subsystems may evolve at different or changing rates
• dynamic identification of weak interactions within a complex system and use of such information to represent the system in a factored representation for efficient approximate inference
• dynamic allocation of reasoning resources to focus reasoning on subsystems in critical state or subsystems that are about to undergo major discrete events
• formal characterization of the approximation error induced by factored approximations to Bayesian updating

Education

• Ph.D. in Computer Science, June 2006, Harvard University
• M.S. in Computer Science, March 2002, Harvard University
• B.S. in Computer Science, June 2000, Yale University
• B.S. in Electrical Engineering, June 2000, Yale University

Publications

• Adaptive Dynamic Bayesian Networks
  Brenda Ng
  2007 Proceedings of the American Statistical Association, Section on Bayesian Statistical Science
  Joint Statistical Meetings, Salt Lake City, July 2007.
• Global/Local Dynamic Models
  Avi Pfeffer, Subrata Das, David Lawless and Brenda Ng
  Proceedings of the 20th International Joint Conference on AI, Hyderabad, India, January 2007.
• Factored Inference for Efficient Reasoning of Complex Dynamic Systems
  Brenda Ng, Ph.D. Dissertation, Harvard University, May 2006.
• Continuous Time Particle Filtering
  Brenda Ng, Avi Pfeffer and Richard Dearden
  Proceedings of the 19th International Joint Conference on AI, Edinburgh, United Kingdom, August 2005.
• Factored Particle Filtering for Data Fusion and Situation Assessment in Urban Environments
  Subrata Das, David Lawless, Brenda Ng and Avi Pfeffer
  Eight International Conference on Information Fusion, Philadelphia, July 2005.
• Factored Sampling for Efficient Tracking of Large Hybrid Systems
  Brenda Ng, Avi Pfeffer and Richard Dearden
  Technical report TR-03-05, Harvard University, March 2005.
• Incremental thin junction trees for Dynamic Bayesian Networks
  Frank Hutter, Brenda Ng and Richard Dearden
  Technical report TR-AIDA-04-01, Darmstadt University of Technology, March 2004.
• Factored Sampling for Hierarchical Monitoring of Complex Hybrid Systems (an extended abstract)
  Brenda Ng and Avi Pfeffer
  Proceedings of the 1st Robosphere Workshop, Moffett Field, November 2002.
• Factored particles for scalable monitoring
  Brenda Ng, Leonid Peshkin and Avi Pfeffer
  Proceedings of the 18th Annual Conference on Uncertainty in AI, Edmonton, Canada, July 2002.

Teaching Experience

• Probabilistic Reasoning (CS282 / Pfeffer / Fall'05)
• Decision Theory (ES201 / Schick / Fall'02)
• Principles of Programming Languages (CS152 / Ramsey / Fall'00)