ALESSANDRO PANELLA
Presentation on Bayesian Nonparametrics
I gave a talk on nonparametric Bayesian methods at the Computer Science Department's Machine Learning seminar in the Spring of 2013. The talk includes an introduction on Bayesian learning and focuses on the Dirichlet process, introduced first in the context of mixture models, and then from a more theoretical perspective. Take a look at the slides below, or download them.
Presentation on Probabilistic Reasoning
I gave a presentation about probabilistic representation and reasoning for the Knowledge Representation reading group at the UIC CS Department. The talk was about different types of uncertain information and Bayesian networks. I hope this talk will be followed by another about First-Order Probabilistic Languages, probably in the Fall 2010 semester. The slides of the presentation can be found here.
CS511 - A sample solution to HW4
This is a sample solution to assignment 4 for CS501 (Artificial Intelligence 2). I have tried to consider a general case. To students: your expected solution didn't necessarily had to be at this level of detail and different assumptions are possible.
You can download the documentation with the report of the results here.
I implemented the Influence Diagrams in Netica and run tests using the Netica Java API within an Eclipse project. You can download the source (both the .neta networks and the Java code) here.
Please note that the code is poorly commented and not very well structured. I wrote it quite fast. I plan on uploading a reviewed version of the code.