Compression can have many variants - encryption, encoding, storage
pre-processing, information distillation, extraction and assimilation,
lossily (sp?) and losslessly all in different levels and durations of
preservation.  If an AGI is receiving data from 1000 video camera's running
full time it can't losslessly store all that data so it has to extract
meaningful information from it when it can or needs to, it may losslessly
compress some of it, lossily some, and then process and extract information
or target it for future processing in stages as storage availability
dictates.  It would need to classify regions of data with a level of
importance and throw certain levels away as time goes on and perhaps
recompress more lossily later depending on circumstances.  It's basically a
real-time adaptation to the finiteness of contemporary storage environments
compared to data consumption rates required by the particular AGI
implementation.  Human brains do the same thing, we are continuously
"compressing" lossily and recompressing and throwing away or reassigning to
higher latency storage regions data that becomes less relevant.  It's a
lossy compression that our brains do, from a generalized standpoint.  But
you COULD make an AGI that deals only with lossless compression, meaning it
either stores the data or doesn't.

John

-----Original Message-----
From: Matt Mahoney [mailto:[EMAIL PROTECTED] 

>> all this talk of compression tests is just so much fantasy.
>> 
>> But, to reiterate, I don't need those further tests to make my case: 
>> the mere fact that a perfectly functional AGI system would be classified 
>> by your test as being not an AGI is enough to make the test a failure.
>> 
>> 
>> Richard Loosemore.

> Isn't the fact that most of the available computing power on the Internet
is
> running on deterministic hardware enough reason to implement a
deterministic
> AGI.  Or do you have an analog computer in mind, like growing brains in
> bottles or something?

>-- Matt Mahoney, [EMAIL PROTECTED]


-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=231415&user_secret=fabd7936

Reply via email to