|
Areas of interest: machine learning, graphical models, signal processing, information
theory, compressed sensing, convex optimization, interpretability; dynamical systems,
functional data analysis / spatial statistics / shape analysis
Models of interest: Independent
Component/Subspace Analysis, Mixture of Kalman Filters; network
models; topic models; differential equation models; Non-Parametric
Bayes.
Neat ideas: geometry of exponential families, automatic optimization, manifold learning, information bottleneck method, probabilistic programming languages
"Follow Occam street, but do not
stop!" -here
tutorials
- Introduction to Kolmogorov Complexity (with Liliana Salvador) (slides), 45 minutes.
- Introduction to Machine Learning and Bayesian inference (slides), 45 minutes.
video demos
slice sampling
general-purpose code
R: R-helpers
Julia: B-Splines
|
|
S. Carré, F. Gabriel, C. Hongler, G. Lacerda, G. Capano - Smart Proofs via Smart Contracts: Succinct and Informative Mathematical Derivations via Decentralized Markets
G. Lacerda - Identification of gene modules using a generative model for relational data - UBC Master's thesis (2010), supervised by Jennifer Bryan.
G. Lacerda, P. Spirtes, J. Ramsey, P.O. Hoyer - Discovering Cyclic Causal Models by ICA (UAI2008), video lecture with slides) extends LiNGAM to discover cyclic models; The non-Gaussian model leads to a finer level of identifiability than what can be achieved in the Gaussian case (e.g. by Richardson's CCD), and allows us to relax the faithfulness assumption.
P. O. Hoyer, A. Hyvärinen, R. Scheines, P. Spirtes, J. Ramsey, G. Lacerda, and S. Shimizu - Causal discovery of linear acyclic models with arbitrary distributions Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence (UAI-2008)
How to intelligently combine LiNGAM with methods based on conditional independence tests. This is useful when it may be the case that more than one, but not all error terms are Gaussian.
G. Lacerda - Upper-Bounding Proof Length with the Busy Beaver (2008) - I derive an (uncomputable) upper bound on the length of the shortest proof of any given statement, as a function of the length of the statement; and briefly discuss implications. Mathematically trivial, but apparently original.
N. Matsuda, W. Cohen, J. Sewall, G. Lacerda, and K. R. Koedinger (2008) - Why tutored problem solving may be better than example study: Theoretical implications from a simulated-student study. In Proceedings of the International Conference on Intelligent Tutoring Systems.
N. Matsuda, W. Cohen, J. Sewall, G. Lacerda, and K. R. Koedinger (2007) - Predicting students performance with SimStudent that learns cognitive skills from observation. In R. Luckin, K. R. Koedinger & J. Greer (Eds.), Proceedings of the international conference on Artificial Intelligence in Education (pp. 467-476). Amsterdam, Netherlands: IOS Press.
N. Matsuda, W. Cohen, J. Sewall, G. Lacerda, and K. R. Koedinger - Evaluating a Simulated Student using Real Students Data for Training and Testing, In C. Conati, K. McCoy & G. Paliouras (Eds.), Proceedings of the international conference on User Modeling (LNAI 4511) (pp. 107-116). Berlin, Heidelberg: Springer.
S. F. Adafre, W. R. van Hage, J. Kamps, G. Lacerda, and M. de Rijke - The University of Amsterdam at CLEF 2004, In: C. Peters and F. Borri, editors, Working Notes for the CLEF 2004 Workshop, pages 91-98, 2004.
|
|