November 17th, 2008

teunc

Lutefisk and Lutefizz

Poll #1299627 Lutefisk / Lutefizz poll

Do you know what the following are?

lutefisk
13(86.7%)
lutefizz
0(0.0%)

Have you ever HAD the following?

lutefisk
2(66.7%)
lutefizz
1(33.3%)
lutefizz
0(0.0%)

Do you ever plan to have the following

lutefisk, maybe
3(42.9%)
lutefisk, definitely
2(28.6%)
lutefizz
0(0.0%)
lutefizz
0(0.0%)
lutefizz
0(0.0%)
lutefizz
0(0.0%)

Why not lutefisk?

Because.
7(100.0%)

Anything else to add?



--
Banazir
  • Current Mood
    silly silly
bayesian

Back to Bayesics

To anyone who's been waiting for me to post about Bayesian neworks, inference, and learning again:

It hasn't escaped my notice that bayesnets has been dormant, not least of which is because Bayesian Network tools in Java (BNJ) development has been on hiatus since 2006. This is due partly to zengeneral's venturing forth to found his startup, and more because we have been working to get our comptranslation infrastructure developed.

However, if you have an interest in Bayes nets or BNJ, please do feel free to comment on this post, and I'll definitely reply this week if not today.

ETA, 08:20 CST Tue 18 Nov 2008: That goes doubly so for our information extraction infrastructure.
--
Banazir
biology

Evolutionary Computation: A Cautionary Tale

True story:

In 1999, I wrote a paper for a workshop on Data Mining with Evolutionary Algorithms (DMEA) at the the Genetic and Evolutionary Computation Conference (GECCO) and the American Association for Artificial Intelligence (AAAI). When it was accepted, I notified a co-worker who had employed two of the students who had worked on the paper with us, co-supervised them, and provided some of the test bed data from a data mining problem (classification of auto insurance policies by expected paid loss).

On April 7, 2000, a year after the GECCO-AAAI DMEA paper was accepted and about nine months after it was published, I was contacted via e-mail by the co-worker with Collapse )

I replied Collapse )

I was asked to tell the publishers Collapse )

Collapse )

I will say this for the record:

I teach evolutionary computation, a branch of computer science that studies "evolutionary techniques inspired by mechanisms from biological evolution such as natural selection, mutation and recombination to find an optimal configuration for a specific system within specific constraints" (Wikipedia, 2004).

Come and get me.
I dare you.


--
Banazir