Jason (jcreed) wrote,

Here's an interesting but ambitious sounding paper by lincoln3's hero, professor Selmer Bringsjord from RPI. Thanks to gustavolacerda for the link. I haven't finished reading it yet, but I hope it doesn't depend on the "soap film process" for actually being a successful P=NP oracle even in a weakened sense. It seems self-evident to me that "pulling the structure out of the soap film and waiting for it to settle" is essentially a gradient-ascent algorithm going on, which by no means are we prohibited from implementing anyway (I imagine it's a ubiquitous primitive technique in approximation algorithms, maybe someone with more complexity theory knowledge than me can back me up here) and which by no means is certain to actually be correct algorithmically.

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded