Posted by
Owen Densmore on
Oct 13, 2008; 5:22pm
URL: http://friam.383.s1.nabble.com/Main-Page-Statistical-Mechanics-Algorithms-and-Computations-tp1307184p1329592.html
Well, winter is besetting us, so it occurred to me that we might want
to turn either the Krauth book (the subject), or your earlier
excellent find:
Information Theory, Inference & Learning Algorithms
David J. C. MacKay
.. into a group reading at the sfComplex. Our Data Mining one was
interesting.
I decided I went at the Krauth book with the wrong mind set, so
started over, looking at it as a conversation with an expert delighted
to give a deep and complete look at the subject. This has led me to
write some simple netlogo example programs, looking at several
distributions used in simple Monte Carlo implementations. Its really
kinda fun! Also a bit embarrassing when I come up with distributions
that are a bit unexpected. I think this area takes a *lot* of care!
I gotta say that Krauth hits on a lot of topics heard in the halls of
SFI.
MacKay's book is quite deep and broad as well, and has the advantage
of being available as a PDF. I haven't looked at his site recently,
but he also had several open source implementations of interest. I
went after his first chapter with the J programming language for the
hell of it (J is an APL derivative, also by Iverson .. both Ken and
his son). This was the one where Dilbert was used as a source for
noisy transmission lines. I bet most of it too could be netlogo-
ized. Or possibly R or Sage.
I ramble .. but .. would some of us be interested in A Winter's Read
in Mathematics?? I can bring the book to wedtech or other venues.
Like beer. Just for instance.
-- Owen
On Oct 11, 2008, at 11:47 AM, Owen Densmore wrote:
> On Oct 11, 2008, at 5:12 AM, Robert Holmes wrote:
>
>> Owen - how's the book? I've been thinking about brushing up on my
>> statistical mechanics - Robert
>
> I'm just getting started. My first observation is that its not your
> basic textbook. By that I mean it does not start with axioms and
> grind out theorems. This is nice in the sense of being informal,
> but it also leaves a feeling of being a bit "loose".
>
> But it has the advantage of having a good scope: hits nearly all of
> the buzz words in an intense SFI rap!
>
> So I think I will like it, but will look up lots of stuff in
> wikipedia or other like sources to fill in gaps.
>
> One point: I got it considerably cheaper via getting it used: $29.50
> + shipping. I can bring it to the next wedtech if you'd like.
>
> -- Owen
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at
http://www.friam.org============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at
http://www.friam.org