[ExI] Immeasurable hubris
anders at aleph.se
Sun Sep 7 18:46:05 UTC 2014
Tomaz Kristan <protokol2020 at gmail.com> , 7/9/2014 3:48 PM:
10 years ago I contemplated and co-wrote an algorithm evolver, you can download it here:
It can do quite non-trivial things.
Looks neat. Have you written up something about it? The demonstrations look good (love those packings) but it would be nice to be able to read how it works.
Overall, genetic approaches to optimization are nice. But they mainly work when the fitness landscape is somewhat forgiving. That is why optimizing algorithms themselves is so tricky, and every little step forward is worth investigating deeply.
Anders Sandberg, Future of Humanity Institute Philosophy Faculty of Oxford University
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat