Scott's Old Ph.D. Research Stuff

I'm currently an independent software engineer working on projects of my own devising (and am not seeking employment). Here's some researchy machine learning stuff I worked on during my Ph.D. at CMU, though.
  • Fast Factored Density Estimation and Compression with Bayesian Networks, by Scott Davies. Ph.D. thesis, 2002. (Tech report CMU-CS-02-138.) PDF gzip'd PostScript

  • Interpolating Conditional Density Trees, by Scott Davies and Andrew Moore. In UAI 2002. (Condensed version; see thesis above for details.) PDF PostScript

  • Mix-nets: Factored Mixtures of Gaussians in Bayesian Networks with Mixed Continuous and Discrete Variables, by Scott Davies and Andrew Moore.

  • Statistical Machine Learning for Large-Scale Optimization. Contributing author; editors J. A. Boyan, W. L. Buntine, and A. Jagota. In Neural Computing Surveys 3, 2000. PDF gzip'd PostScript

  • Bayesian Networks for Lossless Dataset Compression, by Scott Davies and Andrew Moore. In KDD-99. PDF PostScript

  • Applying Online Search Techniques to Continuous-State Reinforcement Learning, by Scott Davies, Andrew Ng, and Andrew Moore. In Proceedings of the Fifteenth National Conference on Artificial Intelligence (AAAI-98), 1998. PDF gzip'd PostScript PostScript

  • Fast Probabilistic Modeling for Combinatorial Optimization, by Shumeet Baluja and Scott Davies. In Proceedings of the Fifteenth National Conference on Artificial Intelligence (AAAI-98), 1998. PDF gzip'd PostScript PostScript

  • Pool-Wise Crossover in Genetic Algorithms: An Information-Theoretic Perspective, by Shumeet Baluja and Scott Davies. In Proceedings of FOGA-98, 1998. PostScript

  • Using Optimal Dependency-Trees for Combinatorial Optimization: Learning the Structure of the Search Space, by Shumeet Baluja and Scott Davies. In Proceedings of the 1997 International Conference on Machine Learning, 1997. PostScript

  • Multidimensional Triangulation and Interpolation for Reinforcement Learning, by Scott Davies. In Advances in Neural Information Processing Systems 9, 1996. PDF PostScript

  • NP-completeness of searches for smallest possible feature sets, by Scott Davies and Stuart Russell. In Proceedings of the AAAI Fall Symposium on Relevance (New Orleans, 1994). PDF PostScript

Here's a link back to Scott Davies's home page.