MATRIX@home

MATRIX@home is a scientific experiment that uses Internet-connected computers to run a sophisticated artificial intelligence system. You can participate by running a free program that downloads and analyzes questions posted to Forum 2000.

It is easy, it is free, and it is fun.

Just click on the Activate! button above.

This is a special early ALPHA release. Please bear with us while we troubleshoot the problems. We are working frantically on a better, graphical Java-powered client that can be used as a screen-saver.

It doesn't work?

Make sure you are using a Java-enabled browser (JDK 1.1 required). Also note that some firewalls will prevent the client from successfully contacting the server.

If you have any questions, please contact us.

What is the client doing?

The client downloads a recent question from Forum 2000 and estimates the neural variance for a small portion of the search space. It then reports the estimated variance back to the server and downloads the next question.

What are all those K's and S's?

They are the two basic combinators of the untyped lambda calculus. All computations are done in the untyped lambda calculus, via a translation of the text through the Lambek calculus.

The two combinators satisfy the basic equations:

K x y = x
S x y z = (x z)(y z)
For example, S K K is the identity function because

S K K x = (K x)(K x) = x
In fact, any Turing computable function can be represented as a combination of the two basic combinators S and K. We have asked Andrej Bauer, the theoretical father of SOMADs, to write an expository note on the untyped lambda calculus. Come back soon and read more about this fascinating topic!