[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: coding and nnet's

From: [email protected] 13-NOV-1995 02:25:05.47:

At 11:40 PM 11/11/95, Michael Pierson wrote:

>I wouldn't be surprised to learn of somebody like FinCEN using
>neural-net systems to do pattern analysis on funds transfers and
>the like, or the NRO or NSA investing research money into
>investigating the usefulness of NNs for image processing or for
>scanning raw ELIN SIGINT or COMINT data.  In fact, I'd be quite
>surprised if they weren't.


I recall reading of some contracts let out, and other RFCs, for AI
programs for FinCEN use. Not surprising. (If I ran FinCEN, this is
what I'd surely have a bunch of folks working on. Scary thought.)

     I find it interesting that FinCEN has a net link (at
http://www.ustreas.gov/treasury/bureaus/fincen/fincen.html). I have
heard about an IRS cash-tracking center in Detroit, where all those
$100+ cash deposits/withdrawals get sent. They probably use similar
setups for analysis, although with more of an individual focus.
     I have recently been contemplating how a private digital cash
system (either national-currency based or private-currency based)
would show up on such an analysis process. My suspicion is that it
would do so via fewer known non-cash transfers. Cash transactions
might actually increase, due to conversions between digital cash
and normal cash. On the other hand, many of those using a private
digital cash scheme may (like myself) tend to use mostly cash
     I am interested in hearing the thoughts of others on this
subject. Particularly useful would be methods to stealth or spoof
such a system; i.e., to either hide a private digital cash system
from it, or to give it so many false positives that they couldn't
distinguish what is real and what is not. I suggest that the latter
would be easier (and would have other valuable applications).

P.S. Does anyone know what is up with the extropians list? Down
again, I suspect.