My ExperimentSuite finally got a proper documentation. For now, you can access it here. It will also become a part of the github repository, where ExperimentSuite is hosted.

Here's the PyBrain video presentation from the ICML 2010 workshop on Open Source Machine Learning Software with installation tutorial and some use cases.

On my publications list, BibTeX entries for most papers are now also available. Feel free to use them if you want to cite any of my papers.

Three journal papers have recently been accepted and are being published. Check out the publications page for more details.

Next semster, I offer a proseminar about Game Theory. We will cover both the classical game theory topics (zero-sum / non-zero sum games, nash equilibrium, minimax algorithm, ...) and some other, game-related topics (altruism, chess computers, back-gammon playing, reinforcement learning, ...). The seminar is full now, but here is the website with the topics. If you are interested, you're welcome to attend the presentations.

Today we announced the release of the latest version of our machine learning library in Python. PyBrain 0.3 is available for download on the website www.pybrain.org or directly on GitHub, where we host the project.

Similar to the ingenius idea of spamgourmet, I programmed my own UCE (unsolicited commercial email) protection feature. Email addresses can be created on the fly (without setup), using a different keyword in front of the @ for a particular site, followed by a dot and the limit, like facebook.5@... Emails arriving at this address are collected by a catchall mailbox and then forwarded to my real email address, or not, if a certain limit (5 in this example) has been reached for this keyword.

My Python script to run experiments and collect data is now available in my github account. It is very generic and works with configuration files. You can run your experiments, kill them in between and continue where you left off. It is multi-core aware (you need python2.6 for this) and can run combinations of parameter ranges for you. All the results are logged in files and folders, and there is a little API that lets you quickly collect and analyse the data. If you want to check it out, just clone it and play around with it. Here's the link: http://github.com/rueckstiess/expsuite/

I finished my snippets section, a collection of syntax-highlighted code fragments and tutorials. It contains both permanent ones (that others might be interested in as well) and temporary snippets, which are a good way of quickly exchanging clipboard contents over the web.