http://research.ibm.com/ibm-q/quantum-card-test/
It kind of renders complex information less probabilistic...
Takes out prophets! π
On 10.08.2019 18:13, Ben Goertzel wrote:
The point is, Matt, you can't copy my quantum predictor without me
knowing you were copying it. Basic principle of quantum
cryptography.
This is irrelevant to AGI though, just a sorta fun thought experiment...
On Sat, Aug 10, 2019 at 10:27 AM Matt Mahoney <[email protected]> wrote:
Evolution is not time reversible so it can't run on a quantum computer. Quantum
processes can also produce uncomputable sequences since they can produce
infinite random bits.
But that aside, let's say you have a simple quantum learner that can predict
any quantum computable sequence, which is any sequence that can be produced on
a Turing machine with a source of random bits (because a quantum computer can
be otherwise simulated classically). Then I can still produce a simple sequence
that you can't predict any better than random guessing. I run a copy of your
quantum predictor, which produced an output from the same distribution, and
return a different symbol. Either the distribution is uniform and you are
guessing, or it's not uniform and you will do worse than guessing.
On Fri, Aug 9, 2019, 9:26 PM Ben Goertzel <[email protected]> wrote:
What if my program was created by quantum evolutionary learning, and
carries out its predictions while running in an uncollapsed quantum
state, coupled with the classical system reading-out its predictions
in a way that doesn't collapse its internal memory states...
Then I can set it up so you can't measure what algorithm my program is
running *without collapsing the state, which I could notice* -- and
even if you emulated my process of quantum evolutionary learning, you
couldn't tell what random program it had produced for me.
So your approach doesn't work for quantum computers... but our
physical universe is a quantum system...
-- Ben
On Sat, Aug 10, 2019 at 9:08 AM Matt Mahoney <[email protected]> wrote:
Suppose you have a simple learner that can predict any computable sequence of
symbols with some probability at least as good as random guessing. Then I can
create a simple sequence that your predictor will get wrong 100% of the time.
My program runs a copy of your program and outputs something different from
your guess.
All the empirical evidence supports this. Good compressors have a lot of code
to handle lots of special cases.
On Fri, Aug 9, 2019, 8:15 PM Ben Goertzel <[email protected]> wrote:
Legg proved there is no such thing as a simple, universal learner. So we can
stop looking for one.
To be clear, these algorithmic information theory results don't show there is
no such thing as a simple learner that is universal in our physical universe...
I'm not saying there necessarily is one, just pointing out that the math is not
so practically applicable as your statement implies...
Artificial General Intelligence List / AGI / see discussions + participants +
delivery options Permalink
--
Ben Goertzel, PhD
http://goertzel.org
βThe only people for me are the mad ones, the ones who are mad to
live, mad to talk, mad to be saved, desirous of everything at the same
time, the ones who never yawn or say a commonplace thing, but burn,
burn, burn like fabulous yellow roman candles exploding like spiders
across the stars.β -- Jack Kerouac
Artificial General Intelligence List / AGI / see discussions + participants +
delivery options Permalink
------------------------------------------
Artificial General Intelligence List: AGI
Permalink:
https://agi.topicbox.com/groups/agi/T1ff21f8b11c8c9ae-Mbd3506af4e5668e440ce29b8
Delivery options: https://agi.topicbox.com/groups/agi/subscription
