Archive of Dmitry Rusakov's Projects and Works

My Ph.D. thesis and related material
- All research during my Ph.D. studies was performed under guidance of Assoc. Prof. Dan Geiger.
- Ph.D. Thesis:
  - Thesis draft (ps, 161 pages, 22 figures, 44MB). (March 17th, 2004.)
  - Pre-candidate exam research report. (November 2000.)
  - Original Ph.D. thesis proposal. (December 1999.)
- Published Papers:
  - Dmitry Rusakov and Dan Geiger "Asymptotic Model Selection for Naive Bayesian Networks" (ps,34MB). Accepted to Journal of Machine Learning Research . (Extended version of UAI-02 paper.)
  - Dmitry Rusakov and Dan Geiger "Automated Analytic Asymptotic Evaluation of Marginal Likelihood for Latent Models". In Proceedings of Uncertainty in Artificial Intelligence-2003 conference.
  - Dmitry Rusakov and Dan Geiger "Asymptotic Model Selection for Naive Bayesian Networks". In Proceedings of Uncertainty in Artificial Intelligence-2002 conference.
  - Dmitry Rusakov and Dan Geiger On Parameter Priors for Discrete DAG Models. In Proceedings of AI and Statistics 2001 conference.
- Additional Research Papers and Technical Reports:
  - Maayan Fishelson, Dmitry Rusakov and Dan Geiger "A Bayesian LOD Score for Linkage Analysis of Complex Diseases". Submitted to Intelligent Systems for Molecular Biology (ISMB) 2004 conference.
  - Technical Report CIS-2000-08 (extended version of AISTATS-01 paper). Matlab code used in the paper is available here.
  - A Tutorial on Schwarz' 1978 paper - research note explaining the details and mathematical machinery behind the standard BIC score, introduced by Schwarz (Annals of Statistics, 1978).
  - Advanced Classical Marginal Likelihood Approximations - this research note summarizes known approximations to marginal likelihoods of data given different Bayesian network models and derives a number of non-standard (non-BIC) asymptotic approximations of marginal likelihood integrals using classical asymptotic analysis (without the use of Watanabe's results).
- Presentation Slides:
  - Seminar slides - presented at the Technion on January 21st, 2004.
  - UAI-03 poster.
  - UAI-02 presentation slides.
  - Candidate exam slides..
  - AISTATS-01 poster.
- Research Projects (described in "Automated Analytic Asymptotic Evaluation of Marginal Likelihood for Latent Models"):
  - BN-dimension project. Evaluates effective dimensionality of arbitrary Bayesian network.
  - Resolution project. Evaluates approximation to the marginal likelihood for "hard" types of log-likelihood function by resolution of singularities of set of maximum likelihood parameters.

Additional research performed during my Ph.D. studies
- Decomposition of Event Sequences into Independent Components. Joint work with Heikki Mannila while staying with Nokia Research. Presented on First SIAM Data Mining Conference. Long version is available.

My M.Sc. thesis and related material
- Selective Sampling for Nearest Neighbor Classifiers - journal paper published by Machine Learning Journal, February 2004, vol 54., issue 2, pages 125-152. This is a joint work with Michael Lindenbaum and Shaul Markovitch.
- Selective Sampling for Nearest Neighbor Classifiers - conference paper, presented on AAAI-99 conference.
- Thesis in English. Hebrew part is also available. The research was conducted under supervision of Dr. Shaul Markovitch and Assoc. Prof. Michael Lindenbaum.
- active_learning.tar.gz- Active Learning of Pixel Classification project (this is my pre-thesis project). The report is available and you can also read my thesis proposal.

Projects done in the spare time
- big_numbers.zip - C++ class that supports natural numbers of arbitrary size, with full support for all arithmetic operations.

Projects done during my M.Sc. studies
- color.ps - Lightness & Color Constancy paper. Final work in Computer Vision course.
- cornell.tar.gz - Cornell box scene model in the MGF modelling language. You can render this scene with RAD radiosity renderer.
- des6-crack.tar.gz - Program for breaking 6-round DES. Very fast and efficient written in C++.
- gauss_filters.tar.gz - a Khoros program for testing oriented gaussian noise filters.
- rote_learning_for_connect-4_game.tar.gz - Rote learning (memorization) for connect-4 game, in LISP. Report is available. Joint work with Ron Bekkerman.
- prunning_for_id3.tar.gz - Prunning Strategies for decision trees build by ID3: C4.5 and IBL, in LISP. Report is available. Joint work with Ron Bekkerman.
- genetic_alg_for_IPD.tar.gz - Genetic algorithm for IPD (Iterated Prisoners Dilemma) game, in C++. Report is available. Joint work with Ron Bekkerman.
- PointLocation.ppt - A PowerPoint presentation of point location algorithm using trapezoidal maps. Joint work with Boris Kronrod.

Projects done during my B.A. studies
- ai.tar.gz - Artificial Intelligence project, in LISP. Last homework in the cs236501 course.
- cg.tar.gz - Computer Graphics project, in C. Last homework in the cs234325 course.
- compilation.tar.gz - Compilation project. Pascal to C interpreter. Joint work with Alex Agranovsky.
- image_rotating_and_warping.tar.gz - Image Transformation project. Command line utility written in C++.
- logical_design_lab.tar.gz - CPU Logical Design Project. Full COMPASS design of 5-stage pipeline RISC CPU.
- raytrace.tar.gz - Ray Tracing project (in C++) and images that we have rendered with it. See images 1, 2 and 3.
- skeletonization.tar.gz - Character Skeletonization project, with X-Motif interface.

Visitor number counter

since last counter reset.

Free counter