On Fri, 10 Jan 2003, Jonathan M. Hollin wrote:
Matt Sergeant wrote:
I've done a lot on bayes for spam (I had an effective bayesian filter before
Paul Graham wrote his article on the subject), but there's not much in it
for a full talk. Maybe a lightning talk. Hmm...
That would be great Matt.
Funny how Paul Graham has become so synonymous with Bayesian
probabilities isn't it? Your work preceeds his, Google Page Rank
apparently uses Bayesians (pre-Graham) and, here in the UK, Autonomy
(http://www.autonomy.com/) have been doing ground-breaking work for a
few years now with Bayesians at the core of their technology. Despite
this, Paul Graham is the first person people think of when you start to
talk Bayesians, I wonder why?
To be fair to Paul Graham, what he did was put it in a language that was
accessible to geeks. Every single paper or book or whatever I've read on
Bayesian probabilities talks in maths gobledegook (although I'm sure Ken
Williams may disagree ;-). Paul Graham made it easy for the layman to
grok and implement.
A bayes lightning talk might be fun though.
--
!-- Matt --
:-get a SMart net/:-
Spam trap - do not mail: [EMAIL PROTECTED]