Research
Work:
My research interests span the areas of optimization, decision analysis and applied statistics. I have a specific interest for quantitative methodologies that manage the risks related to market, environmental or physical uncertainty that is present in industrial and financial decision problems. Applications that have caught my attention include, but are not limited to, portfolio selection, e-commerce, resource allocation, and machine replacement problems (see pdf for more details).
Thesis work
:
Distributionally Robust Optimization in context
of Data-driven Problems. (abstract, pdf)
Publications :
S. Agrawal, E. Delage, M. Peters, Z. Wang, Y. Ye, Unified Framework for Dynamic Pari-Mutuel Information Market Design, accepted to 10th ACM Electronic Commerce Conference. (extended version)
Delage E. Regret-based Online Ranking for a Growing Digital Library, accepted to ACM SIGKDD Conference on Knowledge Discovery and Data Mining 2009. (pdf, extra material)
Delage E., Ye Y. Distributionally Robust Optimization under Moment Uncertainty with Application to Data-Driven Problems, 1st place Nicholson Award 2008, accepted to Operations Research. (pdf)
Delage E., Mannor S. Percentile Optimization in Uncertain MDP with Application to Efficient Exploration, ICML 2007. (pdf)
Delage E., Mannor S. Percentile Optimization for MDP with Parameter Uncertainty, finalist for Nicholson Award 2007, accepted to Operations Research. (pdf,appendix)
Delage E., Lee H. , Ng A Y. A dynamic Bayesian network model for autonomous 3d reconstruction from a single indoor image, CVPR 2006. (pdf, experiments)
Delage E., Lee H., Ng A. Y. Automatic Single-Image 3d Reconstructions of Indoor Manhattan World Scenes, ISRR 2005. (pdf, experiments)
Patents :
Labs J., Delage E.,
Ullman R., Beaton G. (2003) Synchronizing
Method and Apparatus (Patent CA 2433139).
Labs J., Gagnon S.,
Delage E. (2003) Method and apparatus
for estimating frequency offsets for an OFDM burst receiver. (Patent
Pending).
Reports :
Delage, E., Mulye G., Munshi V. (2004) EEG Artifact Cancellation: Adaptive Noise-Cancelling and Component Analysis Based approaches, term paper in a course on Adaptive Signal Processing.
Presentations
:
Delage, E., Distributionally Robust Optimization under
Moment Uncertainty with Application to Data-Driven Problems, presented at
INFORMS annual meeting 2008.
Delage, E., Data-Driven Optimization for Portfolio
Selection, presented at INFORMS annual meeting 2008.
Delage, E., Robust Partitioning for Stochastic Multi-Vehicle
Routing, presented at INFORMS annual meeting 2008.
Delage, E., Percentile
Optimization for Markov Decision Processes with Parameter Uncertainty,
presented at INFORMS annual meeting 2007.
Delage, E., Percentile Optimization in Uncertain MDP with Application to Efficient Exploration. presented at ICML 2007.
Delage, E., A dynamic Bayesian network model for autonomous 3d reconstruction from a single indoor image. Poster presented at CVPR 2006.
Delage, E., Robust Fisher Discriminant Analysis, presented for the McGill-UdeM-MITACS Machine Learning Seminars, McGill University, Winter 2006.
Castro, P. S., Delage, E., The Linear Programming Approach to Approximate Dynamic Programming, presented for the Reasoning and Learning Lab, McGill University, Winter 2006.
Delage, E., Linear Semi-Infinite Programming, presented for the Reasoning and Learning Lab, McGill University, Fall 2005.
Delage, E., A Bayesian Approach For 3D Reconstruction From a Single Image, presented for the Reasoning and Learning Lab, McGill University, Spring 2005.
Delage, E., Linear Programming for POMDP, presented at Stanford for a course on Approximate Dynamic Programming methods, Winter 2005.