Tuesday, March 28, 2006

 

Human competitive competition 2006

This year, the deadline for the application to the human-competitive is 29 May 2006. It will be held as a part of GECCO 2006, and winners will be awarded $10,000. Basically, it's a competition for outstanding results in the GEC field. Eight criteria for the application can be found in Dave's post. Here's another post about the last year winners. Be sure to check the official website if you are interested.

 

EC researchers in India - I

I thought I would start posting about EC researchers in India. I'll start with some of my undergrad professors.

Prof. Nirupam Chakraborti



One of prof. Chakraborti's main research areas is the applications of genetic and evolutionary algorithms to problems is metallurgy and materials science. He first introduced me to genetic algorithms and I've been fascinated by them ever since.

 

New IlliGAL software

We are pleased to announce the release of following software for non-commercial purposes:

χ-ary Extended Compact Genetic Algorithm for Matlab in C++. [Download source] [Abstract] [Documentation in PS] [Documentation in PDF].

χ-ary Extended Compact Genetic Algorithm for in C++. [Download source] [Abstract] [Documentation in PS] [Documentation in PDF].

Extended Compact Genetic Algorithm for in C++. [Download source] [Abstract] [Documentation in PS] [Documentation in PDF].


Other IlliGAL source codes are available here.

Saturday, March 25, 2006

 

New IlliGAL technical reports

IlliGAL is pleased to announce the publication of following technical reports:

Llorà, X. Goldberg, D. E. (2006). The Innovation Pump: Supporting Creative Processes in Collaborative Engineering. IlliGAL Report No. 2006011. [Abstract] [Full paper in PS] [Full paper in PDF].

Lanzi P.L., Loiacono D. (2006). Standard and Averaging Reinforcement Learning in XCS. IlliGAL Report No. 2006010. [Abstract] [Full paper in PS] [Full paper in PDF].

Lanzi P.L., Loiacono D., Wilson S.W., Goldberg D.E. (2006). Classifier Prediction based on Tile Coding. IlliGAL Report No. 2006009. [Abstract] [Full paper in PS] [Full paper in PDF].

Lanzi P.L., Loiacono D., Wilson S.W., Goldberg D.E. (2006). Prediction Update Algorithms for XCSF: RLS, Kalman Filter, and Gain Adaptation. IlliGAL Report No. 2006008. [Abstract] [Full paper in PS] [Full paper in PDF].

Yu, T.-L., Goldberg, D. E. (2006). Conquering Hierarchical Difficulty by Explicit Chunking: Substructural Chromosome Compression. IlliGAL Report No. 2006007. [Abstract] [Full paper in PS] [Full paper in PDF].

Winward, P., Goldberg, D. E. (2006). Fluctuating Crosstalk, Deterministic Noise, and GA Scalability. IlliGAL Report No. 2006006. [Abstract] [Full paper in PS] [Full paper in PDF].

Sastry, K., Johnson, D.D., Thompson, A. L., Goldberg, D. E., Martinez, T. J., Leiding, J., Owens, J. (2006). Multiobjective Genetic Algorithms for Multiscaling Excited State Direct Dynamics in Photochemistry. IlliGAL Report No. 2006005. [Abstract] [Full paper in PS] [Full paper in PDF].

Llorà, X., Sastry, K., Alías, F., Goldberg, D. E., Welge, M. (2006). Analyzing Active Interactive Genetic Algorithms using Visual Analytics. IlliGAL Report No. 2006004. [Abstract] [Full paper in PS] [Full paper in PDF].

Sastry, K., Lima, C. F., Goldberg, D. E. (2006). Evaluation Relaxation Using Substructural Information and Linear Estimation. IlliGAL Report No. 2006003. [Abstract] [Full paper in PS] [Full paper in PDF].

Pelikan, M., Sastry, K., Butz, M.V., Goldberg, D.E. (2006).
Hierarchical BOA on random decomposable problems
. IlliGAL Report No. 2006002. [Abstract] [Full paper in PS] [Full paper in PDF].

Llorà X., Sastry, K. (2006). Fast Rule Matching for Learning Classifier Systems via Vector Instructions. IlliGAL Report No. 2006001. [Abstract] [Full paper in PS] [Full paper in PDF].

Other IlliGAL technical reports and publications are available here.

 

Publish your book blogger!

Blogging has been changing how people disseminate ideas. There are blogs covering almost any topic one may be able to think about. Blogs were seen as mere on-line entities where people express; just a bunch of digital information here and there. But recently they are starting to change into something you can touch and feel.

Blogbinders is a site for turning your blog into a book. For instance, if you have your blog built using Blogger, WordPress, MovableType, Livejournal, or Typepad, Blogbinders allow you to edit it and publish your blog into a book that you can sell. Another interesting alternative is Lulu. It does not directly edit the blog for you, but you can: edit it in a Word file, upload the file to the site, decide the cover and price, and voilá, you have a book. They also shelve it on on-line stores such as Amazon. Lulu is also a very interesting alternative for cheap book publishing. Books that major publishing stream may not be interested in publishing because of their tiny revenue may be published via Lulu-like sites.

And you may think now: will IlliGAL blogging eventually turn into a collectively-written book some day? ;)

Thursday, March 23, 2006

 

March madness and evolutionary bracketology

Jacob Borgerson, an avid sports fan and former IllIGAL member, has developed a GA to predict the bracket for the NCAA men's basketball championship. Given that this is still a work very much in progress the GA results are quite good and from what I know of filling the brackets its no easy task.

However Jacob sent this disclaimer to me: "keep in mind that i'm not the ga-expert that you guys are... i'm probably using an "IN-competent" ga." :). Of course, he is being modest.

Monday, March 20, 2006

 

If Multi-Agent Learning is the Answer, What is the Question?

A colleague of mine has pointed me to what appears an interesting paper discussing the foundation of Multiagent Learning. The paper by Yoav Shoham, Rob Powers, and Trond Grenager from Stanford University has a rather intriguing title "If multi-agent learning is the answer, what is the question?". It will appear on the Journal of Artificial Intelligence.

Monday, March 13, 2006

 

Metadata stores

The DISCUS project has always supported that intuition that annotation capabilities are a must for knowledge and information exchange. For instance, imaging that you are analyzing the KeyGraph generated from a particular discussion (here you can find an example). You may want to enrich such graph with your analysis, comments, or related information. Basically, you want to add metadata to the KeyGraph. If such a capability is available, a whole new bunch of information will need to be efficiently stored to allow, not only fast and easy retrieval, but allow analysis of the added metadata.

The Kowari project is an Open Source, massively scalable, transaction-safe, purpose-built database for the storage, retrieval and analysis of metadata. It provides a simple query language to interact with the metastore (iTQL). If you are familiar with SQL the resemblance will help you get up to speed very fast. The design is oriented to efficiently manage large volume metadata. Informal tests from Joe Frutelle, a NCSA colleague, have convinced me that this metastore can be the way to go for storing the large volumes of metadata that annotation may produce in DISCUS.

Wednesday, March 08, 2006

 

Hunch Engine?

New Scientist this month is to have a report on a GA-based "hunch engine", developed by Icosystem (for whom Eric Bonabeau is chief scientist, apparently). IMHO, New Scientist isn't to be taken as "science" per se, and often contains alot of showey hype, but who knows, this could be interesting.

 

I feel the vibes, man...

PhysOrg is reporting "The World's Fastest Measurements of Molecular Vibrations" which were apparently partially enabled by a GA (if I'm reading this correctly):
The exact resolution of the time measurements is determined by the difference between neighbouring UV frequencies, and is in the range of tenths of a femtosecond. The change over time can be reconstructed from the spectra of two different isotopes. This was done, in the case of the hydrogen experiment, using a complicated genetic computer algorithm.

 

Money, get back...

Just saw a blog item about Tickstation Daily 1.0 which is apparently a database program for "financial instruments", with plotting functions, and this feature:

Automated stock trading strategies can be randomly generated, backtested, and optimized using a novel genetic programming algorithm. Successful strategies can be run on the daily-updated database so that real trading strategies can be implemented. Strategies can be optimized over thousands of stock histories in order to verify the soundness using raw statistics.

Is it a good GA? Will it make us all rich? Who knows? I'm just blogging it, I didn't build it. :-)

 

A buckeye, art, and GAs

Matthew Lewis studies and teaches generative art at Ohio State University and frequently uses genetic algorithms as part of his toolkit. See a short online exhibition of his work here and a list of his projects here. Hat tip generatorx.

 

Irregular blogging

I haven't been a very good blogger this semester. My regular blogging schedule has been upset by my teaching a new prep, an engineering stat course, to 150 students, and the thrice a week lecture preparation and a committee load that would kill a lesser man have sucked the small shards of free time that I used to use to blog out of my schedule.

I'm going to make an effort to get back to blogging a 2-3 times a week over the course of the rest of the semester. Stay tuned.

This page is powered by Blogger. Isn't yours?