David Duvenaud



Curriculum Vitae

My interests lie in the fields of machine learning, Bayesian statistics, and neuroscience. My advisors at Cambridge are Carl Rasmussen and Zoubin Ghahramani . Formerly, I was at UBC, advised by Kevin Murphy.

Publications

Duvenaud, D., Marlin, B., Murphy, K.,  Multiscale Conditional Random Fields for Semi-supervised Labeling and Classification. CRV 2011. [pdf] [demo]

M.Sc. Thesis:  Multiscale Conditional Random Fields for Machine Vision. [pdf]

Duvenaud, D., Eaton, D., Murphy, K. and Schmidt, M. 2009. Causal learning without DAGs.
Journal of Machine Learning Research. [pdf] [code]

I presented the causality work in a poster at the 2009 Machine Learning Summer School in Cambridge. [pdf]

I won the DREAM4 Predictive Signaling Network Modeling Challenge, using a very simple model inspired by the causal learning paper.  Publication forthcoming.







A fun project from my undergrad:





Around 2004 I did some experiments along the lines of Karl Sims' evolved virtual creatures.  My creatures evolved some quite suprising behaviors. Life finds a way! However, I wouldn't recommend the genetic algorithm to anyone for any real optimization task.