Machine Learning Investigations

1. depthMap: learning the motion of spheres from depth maps generated in Unity
2. depthMapReinforcement: learning the motion of spheres as above but using reinforcement learning (very early stage of development)
3. dimensionalityReductionVR: using virtual reality (GearVR) to visualize various methods of dimensionality reduction for MNIST
4. mlProjects.pdf: Ruminations on difficult, speculative but (IMHO) valuable learning projects of special interest to me.
5. unifiedTreeSearch.pdf: A short, terse description of an algorithm for statistical analysis for gambling games, unifying exact inference, random sampling, and Bayesian Monte Carlo Tree Search. (DRAFT IN PROGRESS)
6. Tiny prototype of forward and backward automatic differentation in Scala. https://github.com/mettinger/scala/tree/master/autoDiff