Mailing List Archive

On Wednesday 26 May 2004 05:54 pm, Henry Stern wrote:
> Assuming that my paper is accepted, I will be going. 5 more days of
> nail biting!

Oooh. What was your paper on?

--
Give a man a match, and he'll be warm for a minute, but set him on
fire, and he'll be warm for the rest of his life.

Advanced SPAM filtering software: http://spamassassin.org

Assuming that my paper is accepted, I will be going. 5 more days of
nail biting!

Henry

Theo Van Dinter wrote:

>So is anyone else going to the CEAS (www.ceas.cc) conference in July?
>If there's a "bunch" of us, I'm wondering if we want to get together the
>day before the conference to chat SA or whatever? Just trying to figure
>out what day I should fly in and such. :)
>
>

Matthew Cline wrote:

>On Wednesday 26 May 2004 05:54 pm, Henry Stern wrote:
>
>
>>Assuming that my paper is accepted, I will be going. 5 more days of
>>nail biting!
>>
>>
>
>Oooh. What was your paper on?
>
>
>
Henry Stern, Justin Mason and Michael Shepherd. A linguistics-based
attack on personalised statistical e-mail classifiers.

Abstract:
We present a potential vulnerability of personalised anti-spam filters
where an attacker sends carefully constructed e-mail messages with the
goal of negatively a ecting classi er accuracy. Words from the core of
the English language are randomly "injected" into spam e-mails for the
express purpose of manipulating the probability tables of a naive
Bayesian classifier. This attack method is shown to be successful in
reducing classifier accuracy within a laboratory environment. Barriers
to a real-world implementation and potential countermeasures are discussed.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Henry Stern writes:
> Henry Stern, Justin Mason and Michael Shepherd. A linguistics-based
> attack on personalised statistical e-mail classifiers.
>
> Abstract:
> We present a potential vulnerability of personalised anti-spam filters
> where an attacker sends carefully constructed e-mail messages with the
> goal of negatively a ecting classi er accuracy. Words from the core of
> the English language are randomly "injected" into spam e-mails for the
> express purpose of manipulating the probability tables of a naive
> Bayesian classifier. This attack method is shown to be successful in
> reducing classifier accuracy within a laboratory environment. Barriers
> to a real-world implementation and potential countermeasures are discussed.

'a ecting classi er accuracy'?! back to the drawing board, lads! ;)
(that's "affecting classifier accuracy", obviously.)

- --j.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Exmh CVS

iD8DBQFAtiuuQTcbUG5Y7woRAjGIAJ9aiAD7epDd5iQ8hQsjqvDjSadWcgCeO4TA
v0dQKs1ezRsY2FPdZJGclHU=
=wNw0
-----END PGP SIGNATURE-----

On Thursday 27 May 2004 19:55 CET Justin Mason wrote:
> Henry Stern writes:
> > Abstract:
> > We present a potential vulnerability of personalised anti-spam filters
> > where an attacker sends carefully constructed e-mail messages with the
> > goal of negatively a ecting classi er accuracy. Words from the core of
> > the English language are randomly "injected" into spam e-mails for the
> > express purpose of manipulating the probability tables of a naive
> > Bayesian classifier. This attack method is shown to be successful in
> > reducing classifier accuracy within a laboratory environment. Barriers
> > to a real-world implementation and potential countermeasures are
> > discussed.
>
> 'a ecting classi er accuracy'?! back to the drawing board, lads! ;)
> (that's "affecting classifier accuracy", obviously.)

Oh. I thought it was some intelligent latin proverb I don't know...

Cheers,
Malte

--
[SGT] Simon G. Tatham: "How to Report Bugs Effectively"
<http://www.chiark.greenend.org.uk/~sgtatham/bugs.html>
[ESR] Eric S. Raymond: "How To Ask Questions The Smart Way"
<http://www.catb.org/~esr/faqs/smart-questions.html>

> Oh. I thought it was some intelligent latin proverb I don't know...

It is by the great poet, Adobe Acrobat.

Henry

Henry Stern wrote:
> It is by the great poet, Adobe Acrobat.

I thought it was an attempt to poison the Bayes filters :-)

-- sidney

Justin Mason <jm@jmason.org> wrote:

> 'a ecting classi er accuracy'?! back to the drawing board,
> lads! ;) (that's "affecting classifier accuracy", obviously.)

Someone or something trying and failing to get fancy with "ff"
and "fi" ligatures, apparently.

--
Keith C. Ivey <kcivey@cpcug.org>
Washington, DC