Older blog entries for cerquide (starting at number 26)

Long since last post. Recent news:

I have been to St. Augustine, Florida for the FLAIRS 2003 conference. It was very nice, a mix of local and international that kept a good technical level at a very friendly atmosphere. My presentation was kind of boring and uninteresting. I have to learn how to keep the audience interest.

I have been accepted a paper for the International Conference on Machine Learning to be held in Washington DC in August. I am happy, because this is one of the most prestigious Machine Learning conferences and I have never been to it. I am looking for a cheap travel option to see NY with Elena and then go to Washington to end up relaxing in Punta Cana (that is going to be my honeymoon). Today I have been finishing the experiments for the final version. I was asked by the reviewers to include statistical significance testing for the results. The results are significant with a 95% confidence.

Today I have given a class about "Artificial Intelligence and Machine Learning" in the middle of the "Ronda Universitat" in Barcelona. It was fun.

I am reading Steffen Evers "An Introduction to Open Source Software Development".

I have been asked by two of my students (esteve and angels) to give a class in the middle of the street tomorrow in order to protest against Iraq's invasion. Yes, fxn, two of your future students are also in advogato and you still have not certified them :op

More than one month since my last post.

Things I am actually (or have been lately) involved in :

- Making a presentation for my FLAIRS 2003 paper

- Correcting my student projects.

- Waiting for my PhD thesis presubmission to be accepted

- Learning Ruby, because I will be teaching it on the next course.

- Learning biochemistry and speciallizing in protein folding.

- Reading "Bayesian Theory" by Bernardo and Smith

- Finishing the final version of my PhD thesis

- Saying NO to Iraq war as often as I can

- Trying OpenOffice, which works great for me.

- Psicollogically preparing for my wedding.

- Getting used to UML, Eclipse and EclipseUML.

- Improving my Catalan in order to get the official C-level certificate.

I have programmed a small script to keep my evolution Tasks and Calendar synchronized at home and at work. It uses the Unison file synchronizer. It can be downloaded from here.

And now I have no excuse to keep my tasks and calendar outdated.

After two days with Galeon I have decided to go back to Opera again. Galeon is far too heavy.

I am almost done with my installation. I tested a lot of new software. I am running Gnome at work and home with Sawfish as window manager. Im running Gnome due to its wonderful workgroup suite Evolution from Ximian. It is a very good IMAP mail reader and incorporates a good calendar and task manager. Definitely far ahead of what I am used to see in Unix environments. I am running Galeon as browser, but I have been testing Opera and it does a nice job too. Anyone knows why are browsers so heavy anyhow???

Today I have submitted the final version for our Novatica paper on personalization.

In the next few days I have to finish my PhD thesis summary in order to go through a external review process. I also have to finish my TBMATAN paper to submit it to ICML and to JAIR and to submit the final version of my FLAIRS 2003 paper. Everything before February 15. Ah, by the way I will start teaching Feb, 10!!!

First post from my new job at Universitat de Barcelona. I have spent the morning chating with colleagues trying to get to know them and installing KDE, xemacs and everything else in my new box.

This weekend I have been reading and trying to understand my own developments in order to write Chapter 5 of my thesis and submit the final version of the FLAIRS 2003 paper. Until this morning I was unable to understand that the Indifferent Naive Bayes (proposed by myself!) was based on a different version of the multinomial sample assumption that the one that is commonly used in the Bayesian networks community. I will explain the paper very differently now. The "new" multinomial sample assumption will gain much more relevance in the exposition while the principle of indifference will be not so relevant.

I have been reading about Reference Analysis by Bernardo. There is one thing that puzzles me. How can it be that Bernardo's method does not coincide with the Law of succession in the Bernoulli distribution? This makes me feel reserved about using Reference Analysis in my thesis, also because it is awfully complicated. I think I still feel more inclined to follow Jaynes advice and apply the principle of indifference and maximum entropy.

I have a copy of most of the books I need to prepare the course. I will start on Wednesday.

I have decided to use both Craig Larman's "Applying UML and patterns (2nd ed)" and Fowler's "UML Distilled" as support for the UML part of the course.

For the algorithmic part I am thinking in Goodrich's "Data Structures and Algorithms in Java (Second Edition)" but also in "Data Structures and Problem Solving Using Java (2nd Edition)" by Mark Allen Weiss and in "Data Structures & Algorithms in Java (Mitchell Waite Signature Series)" by Mitchell Waite, Robert Lafore. I will like to read the three of them, but if I am not able to grab them I think I will prefer Goodrich's because of the good support material.