Alexandre Bouchard-Côté's Home Page

General

• Email: bouchard AT cs.berkeley.edu
• Occupation: Fourth year Ph.D. student
• School: UC Berkeley.
• Supervisors: Mike Jordan and Dan Klein.
• English speakers call me Bouchard or Alex
• Mandarin speakers sometimes call me 卜利森
• NLP research group at Berkeley.
• Resumé (last updated: June 09)

Research Interests

My main field of research is statistical machine learning. I am interested in the mathematical side of the subject as well as in the applications to statistical Natural Language Processing (NLP) and computational biology.

I am currently working on high-resolution computational models for evolutionary processes, which can be used to automatically reconstruct proto-languages or to align proteins. This work has the potential to bring methods from computer science to bear on significant problems in both historical linguistics and biology.

Other current or recent research projects include: MCMC, variational inference and using non-parametric Bayesian statistics to train machine translation models.

In the past, I also did some work on logical characterization and approximation of labeled Markov processes and on reinforcement learning.

Other topics of interest include: probability theory, design and analysis of randomized algorithms.

Refereed Publications

• Alexandre Bouchard-Côté and Michael I. Jordan. (2009) Optimization of Structured Mean Field Objectives. Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI09). Montreal, Canada.
  [paper][poster][slides]
• Alexandre Bouchard-Côté, Thomas L. Griffiths and Dan Klein. (2009) Improved Reconstruction of Protolanguage Word Forms. Proceedings of the North American Chapter of the Association for Computational Linguistics (NAACL09). Boulder, USA.
  [paper][slides]
• Alexandre Bouchard-Côté, Michael I. Jordan and Dan Klein. (2009) Efficient Inference in Phylogenetic InDel Trees. Advances in Neural Information Processing Systems 21 (NIPS). Vancouver, Canada.
  [paper][poster][slides]
• John DeNero, Alexandre Bouchard-Côté and Dan Klein. (2008) Sampling Alignment Structure under a Bayesian Translation Model. Proceedings of the 2008 Conference on Empirical Methods on Natural Language Processing (EMNLP08). Waikiki, USA.
  [paper]
• Alexandre Bouchard-Côté, Percy Liang, Thomas Griffiths and Dan Klein. (2008) A Probabilistic Approach to Language Change. Advances in Neural Information Processing Systems 20 (NIPS). Vancouver, Canada.
  [paper][poster]
• Alexandre Bouchard-Côté, Percy Liang, Thomas Griffiths, and Dan Klein. (2007) A Probabilistic Approach to Diachronic Phonology. Proceedings of the 2007 Conference on Empirical Methods on Natural Language Processing (EMNLP07). Prague, Czech Republic.
  [paper][slides][data]
• Percy Liang, Alexandre Bouchard-Côté, Dan Klein, and Ben Taskar. (2006) An End-to-End Discriminative Approach to Machine Translation. Proceedings of the 44th Annual Meeting of the Association for Computational Linguistics (ACL06). Sydney, Australia.
  [paper]
• Alexandre Bouchard-Côté, Norm Ferns, Prakash Panangaden, and Doina Precup. (2005) An Approximation Algorithm for Labelled Markov Processes: towards Realistic Approximation. Proceedings of the International Conference on Quantitative Evaluation of Systems. Torino, Italy.
  [paper][slides][code]

Workshop Papers, Reports and Presentations

• John DeNero and Alexandre Bouchard-Côté. (2009) A Hierarchical Dirichlet Process Prior for a Conditional Model of Phrase Alignment. Workshop on statistical NLP at Advances in Neural Information Processing Systems 21 (NIPS). Whistler, Canada.
• Alexandre Bouchard-Côté (2005). Domaines et languages. Canadian Undergraduate Mathematical Conference (CUMC). Kingston, Canada.
• Alexandre Bouchard-Côté (2004). Sparse distributed memories: some theoretical and empirical results. Report and presentation for the Courtemanche Scholarship (first prize).
  [slides][report][code]
• Alexandre Bouchard-Côté (2004). Reinforcement learning and function approximation. Canadian Undergraduate Mathematical Conference (CUMC). Halifax, Canada.
  [slides]

Old Scribbles

• Final project report and presentations for Stat NLP and machine learning.
• Notes prepared about LMP approximation,
• Report for a term project (for McGill's Machine learning course),
• Source code my experiments in the above report were based on,
• Slides for the same project (also available in .pdf),
• A paper on the history of algebra, focusing on the evolution of polynomial solving methods and Galois theory,
• A java program (with sources) that computes the minimal bisimilar system of a given finite LMP,
• Preliminary document on a specialized data structure used by the SDM architecture,
• C++ code (mostly written by Bohdana Ratitch) implementing SDM,
• An AI for the game knockabout (for an ai undergrad class I took in 2003).
• Various fun documents from pre-college projects: Mars, nano, crmt.