About Me
I am currently a PhD student in Prof. Rina Dechter’s Automated Reasoning group.
Graphical models are a widely used framework for reasoning tasks. I am particularly
interested in search-based approaches to inference augmented with standard
variable elimination techniques.
I have worked on exploring the AND/OR Decision Diagram framework for performing
inference and evaluating its quality on solving high-treewidth problems with
problem specific structures such as determinism and context-specific
independence.
I am currently working on dynamic heuristics for AND/OR Branch
and Bound. This work aims to find a good tradeoff between heuristic
computation and node expansion in AOBB.
As of December 2012, I earned my MS in Computer Science from UC Irvine.
Before UC Irvine, I earned my BS in Computer Science from UC Riverside in 2010. There, I worked as an undergraduate research assistant in R-LAIR (Riverside Lab for Artificial Intelligence Research) with Prof. Christian Shelton.
I worked on the CTBN-RLE code base. Specifically, I implemented the structure learning component, which learns a Bayesian network and continuous time Bayesian network structure from trajectory data. It seamlessly incorporates any of the inference methods available in CTBN-RLE to handle the case of partially observed trajectories. I also implemented mean field variational approximate inference for CTBNs, the algorithm described in Cohn et al. 2009.
| Publications |
|---|
| William Lam and Rina Dechter. Empirical Evaluation of AND/OR Multivalued Decision Diagrams for Inference. In Doctoral Programme of CP 2012, Québec City, QC, Canada, October 2012. [pdf | extended version] |
| E. Busra Celikkaya, Christian R. Shelton, and William Lam. Factored Filtering of Continuous-Time Systems. In Proceedings of UAI 2011, Barcelona, Spain, July 2011. [pdf] |
| Christian R. Shelton, Yu Fan, William Lam, Joon Lee, and Jing Xu. Continuous Time Bayesian Network Reasoning and Learning Engine. In Journal of Machine Learning Research, 11, 1137-1140. [link] |