Bruno,

excuse me if I suggest some circularity in you reply. A "learning machine"
is by def. learning SOMETHING and that SOMETHING comes from its inside, if
we do not specify an 'outside' it may explore (which would not be *learning*,
rather *exploring* - a quite different ballgame - maybe followed by 'and
learning *IT'*).
*The applied (ball)game of 'machine' (substituted for 'learning machine',
excluded per se from the 'exploring' function) reminds me of the puzzle of
my midle-school grandkid: which word is the ONE spelled always incorrectly
in every good dictionary? (My wife found it out, immediately, not me). For
the lucky guessers I allow a Coke on NewYear's Eve at his own expense, of
course.*
It depends on 'machine'. Independent? that. too, has to be explained. Maybe
B&B did.

Your question: "Can a machine find a new thing(?)". I refer to Russell's
"patentable" which I wanted to address: a 'new' ('patentably new'?) thing is
not necessarily a (sorry for the Ger.) "noch nie dagewesen" - it can be not
yet described (but knowable - a new combination of elements usually applied
for different patterns etc.). A good example is in this thread about
"electricity" as NOT describable to a medieval scientist: it might have been
"brand new" and unknown, but it still fits into the 'knowables', so I think
about more 'real' novelty.
E.g. cousins of the Milky Way in outer space before the telescope. That did
not fit into the Flat Earth views. -  A 'better mousetrap' IS 'patentable
and new'.
I agree with your ending: " How to define "new", [for example]. It is a
relative concept."

Happy 2010
John M


On Tue, Dec 29, 2009 at 6:32 AM, Bruno Marchal <marc...@ulb.ac.be> wrote:

>
> On 27 Dec 2009, at 23:16, russell standish wrote:
>
> > On Sun, Dec 27, 2009 at 10:54:53AM -0500, John Mikes wrote:
> >> I wonder if a 'robot' can produce a "noch nie dagewesen" (Ger. for
> >> brand
> >> new) unrelated idea?
> >
> > I do know Hod Lipson from the ALife community, but am not familiar
> > with this particular piece of research. From  the WIRED article, I
> > understand this to be a particular implementation of inductive
> > reasoning by machine. It is impressive enough that this is possible,
> > but I don't for one minute think that they have approached the
> > creative power of a human being. But it certainly feasible that humans
> > are really just more so of what this machine does.
> >
> > Still, the whole area of machine learning, and minimum length
> > description has some very interesting surprises in store, which is why
> > I've never bought Colin's argument. For instance John Koza's genetic
> > programs have created several electronic circuits, some of which were
> > patentable, so fit the requirement of noch nicht dagewesen.
>
>
> And there is the whole computational learning theory which shows that
> machine learners exists.
> Even universal learners exists, but the proofs are necessarily non
> constructive. We cannot recognize such machine even if we are in front
> of them.
> There are a lot of amazing theorems in that field. For example the
> theorem of Blum and Blum, which says that there is something
> infinitely (even non computably) more clever (in learning) than any
> machine: a couple of (independent) machines!
> Learning machines exist, and the theory explains why we cannot build
> them from scratch. Some form of learning-competence can need
> intrinsically long computations/histories, but once there, they can
> multiplied.
>
> Can a machine find a new thing. Of course, from I said above. Can we
> judge if a machine has find something new? This is hard to say. It is
> even hard to judge this in a definitive way with the discoveries made
> by humans. It would need many formal criteria in a place where
> formalization is difficult. How to define "new", for example. It is a
> relative concept.
>
>
> Bruno
> http://iridia.ulb.ac.be/~marchal/
>
>
>
> --
>
> You received this message because you are subscribed to the Google Groups
> "Everything List" group.
> To post to this group, send email to everything-l...@googlegroups.com.
> To unsubscribe from this group, send email to
> everything-list+unsubscr...@googlegroups.com<everything-list%2bunsubscr...@googlegroups.com>
> .
> For more options, visit this group at
> http://groups.google.com/group/everything-list?hl=en.
>
>
>

--

You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-l...@googlegroups.com.
To unsubscribe from this group, send email to 
everything-list+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/everything-list?hl=en.


Reply via email to