Self-driving cars

8 views
Skip to first unread message

Evgenii Rudnyi

unread,
Dec 28, 2011, 3:29:07 PM12/28/11
to everyth...@googlegroups.com
I have attended a class

Introduction to AI
http://www.ai-class.com

and there are two interesting videos from it about self-driving cars:

Part 1
http://www.youtube.com/watch?v=V_BJUBpuvFE

Part 2
http://www.youtube.com/watch?v=kqDvbguZsAA

I have no idea if this are Lobian machines or not (such a term has not
been mentioned in the class, it looks like that engineers unaware of
it). Yet I guess it is a good object to discuss what is perception,
qualia, conscious, etc.

Another question would be if such a car could be considered as an
observer in quantum mechanics.

Evgenii

http://blog.rudnyi.ru

meekerdb

unread,
Dec 28, 2011, 3:57:09 PM12/28/11
to everyth...@googlegroups.com

I don't know whether it's a Lobian machine or not; I guess that depends on its program.
But I'm pretty sure that being a big macroscopic thing with lots of degrees of freedom and
interaction with the environment it will "collapse" wave functions.

Brent

> mechanics.
>
> Evgenii
>
> http://blog.rudnyi.ru
>

Bruno Marchal

unread,
Dec 29, 2011, 10:48:02 AM12/29/11
to everyth...@googlegroups.com

Yes. From its first person point of view. It is a good idea to put
"collapse" in quotes. But I am not sure it needs to be macroscopic for
"collapsing" the wave. It needs to be macroscopic only for making a de-
collapsing impossible in practice. If information leaked to much in
the environment, we can no more erase it, and get back to the initial
pure state.
But if QM is correct, even a black hole cannot erase information, and
that seems to explain some observable black hole feature.

Bruno

http://iridia.ulb.ac.be/~marchal/

Evgenii Rudnyi

unread,
Dec 29, 2011, 12:16:34 PM12/29/11
to everyth...@googlegroups.com
On 29.12.2011 16:48 Bruno Marchal said the following:

>
> On 28 Dec 2011, at 21:57, meekerdb wrote:
>

...

>>> Another question would be if such a car could be considered as an
>>> observer in quantum
>>
>> I don't know whether it's a Lobian machine or not; I guess that
>> depends on its program. But I'm pretty sure that being a big
>> macroscopic thing with lots of degrees of freedom and interaction
>> with the environment it will "collapse" wave functions.
>
> Yes. From its first person point of view. It is a good idea to put
> "collapse" in quotes. But I am not sure it needs to be macroscopic
> for "collapsing" the wave. It needs to be macroscopic only for making

> a de-collapsing impossible in practice. If information leaked to much


> in the environment, we can no more erase it, and get back to the
> initial pure state. But if QM is correct, even a black hole cannot
> erase information, and that seems to explain some observable black
> hole feature.

Well, some time ago Rex posted a quote about the rock:

http://blog.rudnyi.ru/2011/02/rock-and-information.html

In this respect, a question would be what is the difference between a
self-driving car and a rock. Both have a lot of degrees of freedom and
both interact.

Bruno, what do you mean by the first person view? Does a self-driving
car has it? Does a rock has it?

Evgenii
--
http://blog.rudnyi.ru


Bruno Marchal

unread,
Dec 29, 2011, 1:40:49 PM12/29/11
to everyth...@googlegroups.com

In the UDA, a first person point of view is described by the content
of its personal (but still sharable in some sense) diary.
In AUDA the first person discourse is defined by what is both
justifiable and true (it is a notion of rational knowledge, but it can
and must be extended to inferable and true in some self-transforming
context).

So a self-driving car is probably much more close to have a first
person view than a rock, especially if you make it possible for the
car to memorize its short term instances of computation (sensing,
planning, etc.) into a "long scenario involving herself".

The singularity point here will be when self-driving car will drink
alcohol and do crazy irresponsible moves on the road to impress other
cars.

The rock? It is not even an object, with comp, but a stable pattern in
a continuum of computational histories, that you don't have to reify.
*IF* you reify them, you might say that a rock is composed of
infinities of universal machine histories, including all universal
dovetailing, and it that sense the rock instantiate consciousness, but
even in that case, it does not make the perceptual rock object into a
person.

I more or less agree also with Chalmers' conclusions.The physical rock
(if we knew what that could be) is not known to physically implement
arbitrary complex computations. There is a statistical repetition in
rock and crystal which prevent turing universal grow. Consciousness
seems to incarnate person relatively to us only near in-equilibrium.

We might miss something about rock, of course. May be some rock are
alien disguised into rock, but then it is up to you to show me their
personal diaries. Are there still rock also?

Without unifying gravitation and the quantum; we can only say that we
still don't know what a rock is, and there are no evidence, in the
common everyday sense, for seeing a rock like a person (despite the
strong and perhaps deep poetical appeal of the idea).

Some Penrose pavements have a Turing universal grow, and they seem to
hide interesting von Neumann algebra. So I would already be far more
cautious with enough big quasi-crystals, and a long time ago, I
thought understanding that some star might collapse into exotic
quantum state of matter through elaborate internal quantum
computations. So to negate consciousness to some "outside body like
object" is always difficult, but in practice, we are interested only
in the doing or the possibility of sharing experiences with some
other. can we share something with that entity?

And with comp, the situation is worst. That's sharing is all we have.
And we have to derive what matter is from that first person sharing,
and explain what that unification of physical laws is all about, if
even possible, to have an idea of what a rock is.

In fine, the real question is do you accept your daughter marry a
rock? Does rock feels pain and should we create a committee to defend
the right of rocks? Can rock votes? It is more easy to understand that
this might make much more sense for the elaborate self-moving machines
of some futures, than for rocks. ... Self-driving cars rocks :)

Bruno

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

http://iridia.ulb.ac.be/~marchal/

Bruno Marchal

unread,
Dec 29, 2011, 1:56:09 PM12/29/11
to everyth...@googlegroups.com

On 28 Dec 2011, at 21:29, Evgenii Rudnyi wrote:

> I have attended a class
>
> Introduction to AI
> http://www.ai-class.com
>
> and there are two interesting videos from it about self-driving cars:
>
> Part 1
> http://www.youtube.com/watch?v=V_BJUBpuvFE
>
> Part 2
> http://www.youtube.com/watch?v=kqDvbguZsAA

Cute video.


>
> I have no idea if this are Lobian machines or not (such a term has
> not been mentioned in the class, it looks like that engineers
> unaware of it). Yet I guess it is a good object to discuss what is
> perception, qualia, conscious, etc.
>
> Another question would be if such a car could be considered as an
> observer in quantum mechanics.

Yes, with comp, in the Everett formulation of QM. (where observers's
body obey QM). In the others theories, it depends on the theory of
mind in use.

In that theory pure QM theory, measurement is only interaction. This
makes the quantum superposition of the object contagious to the
observer (and thus you get the Many Worlds).

In Everett's theory, observation is interaction. The effect of the
interaction is the memory of the events, which makes sense in the
partial trace of the observable. Of course you need quite an amount of
self-interaction to develop into a human-like observers. Perhaps you
need dinosaurs fights and asteroids to get their exact actual
idiosyncrasies.

Bruno


http://iridia.ulb.ac.be/~marchal/

Charles Goodwin

unread,
Dec 29, 2011, 2:44:34 PM12/29/11
to everyth...@googlegroups.com
Hi Bruno

What observable properties of black holes may be explained by the fact that they don't erase information? Is that a purely hypothetical suggestion, or is it something we may observe in the near future, or may have already observed, indirectly?

Thanks!

Cheers,
Charles

Bruno Marchal

unread,
Dec 30, 2011, 7:13:48 AM12/30/11
to everyth...@googlegroups.com

On 29 Dec 2011, at 20:44, Charles Goodwin wrote:

> Hi Bruno
>
> What observable properties of black holes may be explained by the
> fact that they don't erase information?

Black hole evaporation. I am thinking about some work by Hawking.
But $any* true erasing of information is forbid in any theory where QM
applies universally. Unitary evolution cannot erase information,
although it can hide it and makes it very hard to recompose.


> Is that a purely hypothetical suggestion, or is it something we may
> observe in the near future, or may have already observed, indirectly?

I think some cosmological observations confirm this. It makes QM
necessary to justify the existence and stability of Black Hole.

Happy new year Charles,

Bruno

http://iridia.ulb.ac.be/~marchal/

meekerdb

unread,
Dec 31, 2011, 2:38:46 AM12/31/11
to everyth...@googlegroups.com
On 12/30/2011 4:13 AM, Bruno Marchal wrote:
>
> On 29 Dec 2011, at 20:44, Charles Goodwin wrote:
>
>> Hi Bruno
>>
>> What observable properties of black holes may be explained by the fact that they don't
>> erase information?
>
> Black hole evaporation. I am thinking about some work by Hawking.

Hawking's theory of BH radiation did entail erasing information because the radiation is
purely thermal. That's why it was so controversial. Erasing information would be an
exception the unitary evolution of QM. It is now widely thought that the holographic
principle saves the information on the BH event horizon and Leonard Susskind has an
explanation of how string-theory accomplishes this But it's far from a settled question.

> But $any* true erasing of information is forbid in any theory where QM applies
> universally. Unitary evolution cannot erase information, although it can hide it and
> makes it very hard to recompose.
>
>
>> Is that a purely hypothetical suggestion, or is it something we may observe in the near
>> future, or may have already observed, indirectly?
>
> I think some cosmological observations confirm this. It makes QM necessary to justify
> the existence and stability of Black Hole.

There have been no observations of Hawking radiation from a BH, but there are claims of
observing radiation due event horizons created experimentally:

http://arxiv.org/PS_cache/arxiv/pdf/1009/1009.4634v1.pdf

And a Happy New Year to you too, Bruno.

Brent

>
>

Evgenii Rudnyi

unread,
Dec 31, 2011, 9:07:21 AM12/31/11
to everyth...@googlegroups.com
>So a self-driving car is probably much more close to have a first
>person view than a rock, especially if you make it possible for the
>car to memorize its short term instances of computation (sensing,
>planning, etc.) into a "long scenario involving herself".

Good point. Thanks Bruno. A self-driving car does have an estimate of
its current state and then it updates it both internally and based on
external measurements. It also makes some planning, soft of what to do next.

Yet, if we consider a self-driving car and a rock from the viewpoint of
physicalism (or could be even better atomism), then the difference will
be much more difficult to find. After all there are in both cases
interacting electrons and nuclei (well probably some electromagnetic
waves as well) and nothing more.

Evgenii

On 29.12.2011 19:40 Bruno Marchal said the following:

>> self-driving car has it? Does a rock has it?

meekerdb

unread,
Dec 31, 2011, 4:05:14 PM12/31/11
to everyth...@googlegroups.com
On 12/31/2011 6:07 AM, Evgenii Rudnyi wrote:
> >So a self-driving car is probably much more close to have a first
> >person view than a rock, especially if you make it possible for the
> >car to memorize its short term instances of computation (sensing,
> >planning, etc.) into a "long scenario involving herself".
>
> Good point. Thanks Bruno. A self-driving car does have an estimate of its current state
> and then it updates it both internally and based on external measurements. It also makes
> some planning, soft of what to do next.
>
> Yet, if we consider a self-driving car and a rock from the viewpoint of physicalism (or
> could be even better atomism), then the difference will be much more difficult to find.
> After all there are in both cases interacting electrons and nuclei (well probably some
> electromagnetic waves as well) and nothing more.
>
> Evgenii

That kind of "nothing buttery" applies to all explanations because any real explanation is
in terms of simpler things and relations. You could as well say Bruno's explanation is
nothing but arithmetic.

Brent

Evgenii Rudnyi

unread,
Dec 31, 2011, 4:49:55 PM12/31/11
to everyth...@googlegroups.com
On 31.12.2011 22:05 meekerdb said the following:

Well, I guess, it is still an interesting question. Let us take for
example the next sequence

A rock - a ballcock in a toilet - a PID controller in a thermostat - a
self-driving car

What is the difference from the viewpoint of atomism? If you know some
speculations in this direction, I would really appreciate a link, as
this question puzzles me now.

It well might be that a good answer could help us to understand what for
example supervenience is. Otherwise, it is for me just a buzzword like
emergent properties.

I agree, this concerns Bruno theorem as well. It would be nice to learn
what is the difference along this sequence from Bruno's viewpoint as
well. Well, probably some difference here in this respect is a diary.

Evgenii

meekerdb

unread,
Dec 31, 2011, 5:06:11 PM12/31/11
to everyth...@googlegroups.com

It's in the organization of the atoms. But of course you knew that. But why would you
want to force that constraint on the explanation? Why not from the viewpoint of function,
or the viewpoint of evolution or the viewpoint of economics. My brain isn't equipped to
go from the level of atoms to function. A description of a ballcock in terms of atoms
contains too much extraneous information for something that's defined in terms of it's
function for human purposes.

> If you know some speculations in this direction, I would really appreciate a link, as
> this question puzzles me now.
>
> It well might be that a good answer could help us to understand what for example
> supervenience is. Otherwise, it is for me just a buzzword like emergent properties.

I tend to agree. On the one hand it seems to be a relation of x to y such that Sxy =>
~Sxz for all z=/=y. But y can take different values for the same x at different times.

Brent

Charles Goodwin

unread,
Jan 6, 2012, 7:33:24 PM1/6/12
to Everything List
> Black hole evaporation. I am thinking about some work by Hawking.

Could you point me towards it? I know Hawking conceded a bet on this
recently but I'm not sure why.

> But $any* true erasing of information is forbid in any theory where QM
> applies universally. Unitary evolution cannot erase information,
> although it can hide it and makes it very hard to recompose.

True....but I don't suppose anyone is sure that QM necessarily applies
universally (altho I would bet that it did if I had to!)

> I think some cosmological observations confirm this. It makes QM
> necessary to justify the existence and stability of Black Hole.

More information, please!!! :-)

> Happy new year Charles,

Thank you, and happy new year to you, too!

Best wishes,
Charles

Bruno Marchal

unread,
Jan 7, 2012, 6:30:20 AM1/7/12
to everyth...@googlegroups.com

On 07 Jan 2012, at 01:33, Charles Goodwin wrote:

>> Black hole evaporation. I am thinking about some work by Hawking.
>
> Could you point me towards it? I know Hawking conceded a bet on this
> recently but I'm not sure why.

Search on "Hawking Preskill Thorne bet".
Or on "black hole information paradox 1997".


>
>> But $any* true erasing of information is forbid in any theory where
>> QM
>> applies universally. Unitary evolution cannot erase information,
>> although it can hide it and makes it very hard to recompose.
>
> True....but I don't suppose anyone is sure that QM necessarily applies
> universally (altho I would bet that it did if I had to!)

In science we are never sure. But QM seems to be a very solid
hypothesis. It is the first "exact" theory which has lived longer than
15 years, and indeed after about a century it remains undefeated. By
QM I mean the formulation of QM without the wave collapse (Everett).


>
>> I think some cosmological observations confirm this. It makes QM
>> necessary to justify the existence and stability of Black Hole.
>
> More information, please!!! :-)

Wheeler coined the term "black hole" to make the idea sounding
ridiculous. Indeed Black Hole does not make much sense in classical
general relativity. It would force nature into dividing number by
zero. But quantum mechanics came to rescue them, and today it is
generally admitted that black holes indeed exist, in many forms.
A problem with QM arised: is information falling in a black hole
irreversibly destroyed ? If that is the case, then both classical and
quantum physics are incorrect, if not, it means black hole can
restitute in principle every bit of information which felt in it.
Apparently Hawking bet that the information might be destroyed,
and ... lost the bet. I think you will find many informations on this
by searching on terms like black hole, Hawking, Preskill.
I am not an expert in cosmology (nor physics), so I have only layman
ideas on this. I do have "theological-arithmetical" reason that the
core of physics is highly symmetrical, though, and I do expect that
the time symmetry of nature is something very basic. I explain why in
the paper:

Marchal B., 2005, Theoretical computer science and the natural
sciences, Physics of Life Reviews, Vol. 2 Issue 4 December 2005, pp.
251-289.
http://www.sciencedirect.com/science/article/pii/S1571064505000242

Bruno

http://iridia.ulb.ac.be/~marchal/

meekerdb

unread,
Jan 7, 2012, 4:00:33 PM1/7/12
to everyth...@googlegroups.com
On 1/7/2012 3:30 AM, Bruno Marchal wrote:
>
> On 07 Jan 2012, at 01:33, Charles Goodwin wrote:
>
>>> Black hole evaporation. I am thinking about some work by Hawking.
>>
>> Could you point me towards it? I know Hawking conceded a bet on this
>> recently but I'm not sure why.
>
> Search on "Hawking Preskill Thorne bet".
> Or on "black hole information paradox 1997".
>
>
>>
>>> But $any* true erasing of information is forbid in any theory where QM
>>> applies universally. Unitary evolution cannot erase information,
>>> although it can hide it and makes it very hard to recompose.
>>
>> True....but I don't suppose anyone is sure that QM necessarily applies
>> universally (altho I would bet that it did if I had to!)
>
> In science we are never sure. But QM seems to be a very solid hypothesis. It is the
> first "exact" theory which has lived longer than 15 years, and indeed after about a
> century it remains undefeated. By QM I mean the formulation of QM without the wave
> collapse (Everett).
>
>
>>
>>> I think some cosmological observations confirm this. It makes QM
>>> necessary to justify the existence and stability of Black Hole.
>>
>> More information, please!!! :-)
>
> Wheeler coined the term "black hole" to make the idea sounding ridiculous.

Wheeler coined the term, but not to make it sound ridiculous, rather to popularize the
concept. I understand it had a rather different reference in French than in English.
Perhaps you're thinking of "Big Bang" which was coined by Tommy Gold to mock the concept.

> Indeed Black Hole does not make much sense in classical general relativity. It would
> force nature into dividing number by zero. But quantum mechanics came to rescue them,
> and today it is generally admitted that black holes indeed exist, in many forms.
> A problem with QM arised: is information falling in a black hole irreversibly destroyed
> ? If that is the case, then both classical and quantum physics are incorrect, if not, it
> means black hole can restitute in principle every bit of information which felt in it.
> Apparently Hawking bet that the information might be destroyed, and ... lost the bet.

More accurately, Hawking conceded the bet and now believes that black holes do not destroy
information, but there is no solid evidence or theory to show that. It's just that
Stephen changed his opinion.

Brent

Bruno Marchal

unread,
Jan 7, 2012, 5:44:19 PM1/7/12
to everyth...@googlegroups.com

Gosh you might be right. Sorry.


>
>> Indeed Black Hole does not make much sense in classical general
>> relativity. It would force nature into dividing number by zero. But
>> quantum mechanics came to rescue them, and today it is generally
>> admitted that black holes indeed exist, in many forms.
>> A problem with QM arised: is information falling in a black hole
>> irreversibly destroyed ? If that is the case, then both classical
>> and quantum physics are incorrect, if not, it means black hole can
>> restitute in principle every bit of information which felt in it.
>> Apparently Hawking bet that the information might be destroyed,
>> and ... lost the bet.
>
> More accurately, Hawking conceded the bet and now believes that
> black holes do not destroy information, but there is no solid
> evidence or theory to show that. It's just that Stephen changed his
> opinion.

I share with Preskill the idea that the core fundamental (but not
primitive) physics is QM (without collapse). If the physical universe
is described by a unitary evolution then even a black hole can't
destroy one bit of information. So I do think there is a theory (QM),
even if I think this is an open problem in comp (but the material
hypostases already gives some hint for unitary and other symmetries).
And any evidence that QM applies in the cosmos is evidence that black
hole evolution have to be unitary, or partial trace of unitary
evolution. But I doubt QM makes it possible for information to leak
from one universe to a 'parallel one'. I dunno, I should think about
black hole entanglements, but globally QM evolution is information
preserving. OK?

Bruno

http://iridia.ulb.ac.be/~marchal/

Reply all
Reply to author
Forward
0 new messages