1994-06-20 - Neural nets crack RSA? AAARRGHH!

Header Data

From: rishab@dxm.ernet.in
To: cypherpunks@toad.com
Message Hash: 0a030775649db97f8f41d7048d6c0acdd3740ffe52e2bc1950d716c772808c90
Message ID: <gate.5LP9Nc1w165w@dxm.ernet.in>
Reply To: N/A
UTC Datetime: 1994-06-20 15:33:42 UTC
Raw Date: Mon, 20 Jun 94 08:33:42 PDT

Raw message

From: rishab@dxm.ernet.in
Date: Mon, 20 Jun 94 08:33:42 PDT
To: cypherpunks@toad.com
Subject: Neural nets crack RSA? AAARRGHH!
Message-ID: <gate.5LP9Nc1w165w@dxm.ernet.in>
MIME-Version: 1.0
Content-Type: text/plain


Jim choate <ravage@bga.com>:
> As far as I am concerned if it could be done w/ a neural network, or

Ouch. Neural networks are _deterministic_ (except, possibly, analog VLSI ones
such as Mead's).
Digital neural networks can't do _anything_  that cannot be done by 
algorithms; they run on Turing machines too. Why neural networks are 'magical' 
as Perry says, in certain classes of problems, is that they provide a way to
perform complex algorithmic funtions through a relatively simple interface
to humans.

In any case, how "fuzzy" do you allow your relative magnitude function to be?
Even if a function exists that compares keys with a magnitude difference of
the order of 100 bits, it's useless. For your binary search method to work,
this comparator function would have to be accurate when the difference is on
the order of just 1 bit. 

Yeah, quantum computers ;-)

----------------From owner-cypherpunks  Mon Jun 20 10:33:41 1994
Return-Path: <owner-cypherpunks>
Received: by toad.com id AA20571; Mon, 20 Jun 94 10:33:41 PDT
Received: from src.honeywell.com (moon.src.honeywell.com) by toad.com id AA20548; Mon, 20 Jun 94 10:33:24 PDT
Received: from tbird.src.honeywell.com by src.honeywell.com (4.1/smail2.6.3/SRCv0.25);
	Mon, 20 Jun 94 12:33:26 CDT id AA02136  for cypherpunks@toad.com  at toad.com
Posted-Date: Mon, 20 Jun 94 12:33:15 CDT
Received: by tbird.src.honeywell.com (4.1/SMI-3.2)
	id AA29883; Mon, 20 Jun 94 12:33:15 CDT
Date: Mon, 20 Jun 94 12:33:15 CDT
From: bergstro@src.honeywell.com (Pete Bergstrom)
Message-Id: <9406201733.AA29883@tbird.src.honeywell.com>
To: cypherpunks@toad.com
Subject: re: MAIL: secure mail
Sender: owner-cypherpunks@toad.com
Precedence: bulk

>>>>> On Sat, 18 Jun 1994 21:05:04 -0500 (CDT), Karl Lui Barrus <klbarrus@owlnet.rice.edu> said:

> What I've tried is to do a "|pgp -feat ecarp|rmail erc%khijol@apple.com",
> but pgp just produces empty files!  If I execute it by hand, it works.
> Anyone know either (1) why it isn't working wnd how to fix it, or (2)
> suggect a better alternative?

> The problem may be this: pgp needs a path.

> Try "| PGPPATH=/whatever pgp -feat ecarp | rmail erc%khijol@apple.com"

> This is what I needed to do when I tried something similar with mh
> and slocal.

Keep in mind that if you have a site that follows CERT advisories,
there are usually restrictions on which programs may be used to filter
mail. My site requires a sysadmin to put an entry in a config file to
allow any filter programs to work at all.

One standard filtering program that works well is procmail. This will
typically be "approved" by default and should keep your (Ed's) admin
from having to be involved in your incoming mail.

Pete









Thread