TED talk by Daniel Wolpert

Thought you all might enjoy this TED talk by Daniel Wolpert ...
http://www.ted.com/talks/daniel_wolpert_the_real_reason_for_brains.html

... lots of PCT-relevant arguments (including a role for prediction).

···

________________________________________________________________

Prof ROGER K MOORE BA(Hons) MSc PhD FIOA FISCA MIET

Chair of Spoken Language Processing
Vocal Interactivity Lab (VILab)
Speech & Hearing Research Group (SPandH)
Department of Computer Science, UNIVERSITY OF SHEFFIELD
Regent Court, 211 Portobello, Sheffield, S1 4DP, UK

e-mail:� r.k.moore@dcs.shef.ac.uk
web:� http://www.dcs.shef.ac.uk/~roger/
twitter: @rogerkmoore
Tel: +44 (0) 11422 21807
Fax: +44 (0) 11422 21810
Mob: +44 (0) 7910 073631
Editor-in-Chief: COMPUTER SPEECH AND LANGUAGE
(http://www.journals.elsevier.com/computer-speech-and-language/)
________________________________________________________________

[From Rick Marken (2012.10.05.1530)]

Thought you all might enjoy this TED talk by Daniel Wolpert ...
http://www.ted.com/talks/daniel_wolpert_the_real_reason_for_brains.html

... lots of PCT-relevant arguments (including a role for prediction).

Well, I don't know if "enjoy" is quite the right word to describe my
feeling about the talk (I couldn't make it all the way through,
actually) but it was nice to see that the Very Serious People (VSP) in
robotics are no closer to understanding how things work than do the
VSP in economics (VSP is Krugman's term for the people who are taken
seriously in economics, and shouldn't be).

The talk did have a lot of PCT relevant arguments in the sense that
PCT has shown that all of the arguments are incorrect. I knew things
were going south when Wolpert said that the brain is there to control
movement. He's only off by ~180 degree. The brain, of course, is there
to control perceptions.

It looks like the PCT secret is still safe;-)

Best

Rick

···

On Fri, Oct 5, 2012 at 12:16 PM, R K Moore <r.k.moore@dcs.shef.ac.uk> wrote:

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Rupert Young (2012.10.08.1500 BST)]

It was an interesting video, though do agree with Rick's comments.

However, it is promising that at least the dependency between perception and action is recognised. Hopefully this would mean that PCT would be viewed more favourably.

He seemed to be presenting an explanation far more complicated than it need be. If only he knew the secret! I wonder if he would be open to considering a different paradigm.

Regards,
Rupert Young

Sigh!

Sorry to disappoint Warren, but when you say that you agree with Rick, do
you mean that that you agree ...
  - it's OK to voice an opinion on a subject (i.e. the content of Wolpert's
talk) without fully reviewing the material in question?
  - it's acceptable to publicly question the veracity of someone's
scientific reputation without providing specific reasons?
  - it's productive to express satisfaction that a related field of study
seems not to have all the answers to a bunch of seriously hard problems that
face us all?
  - it's correct to claim that "PCT has shown that all of the arguments are
incorrect" (i.e. Bayesian inference, predictive control, algorithms for
separating 'self' from 'other', action selection etc.) without citing the
relevant peer-reviewed evidence to back it up in each case?

I had hoped that people would see the potential overlaps and opportunities
to move our science and understanding forward collectively.

Roger

···

-----Original Message-----
From: Warren Mansell [mailto:wmansell@gmail.com]
Sent: 06 October 2012 14:28
To: Richard Marken
Cc: r.k.moore@dcs.shef.ac.uk; CSGNET@listserv.illinois.edu
Subject: Re: TED talk by Daniel Wolpert

Agreed Rick! Roger, come on!
W

Sent from my iPhone

On 5 Oct 2012, at 23:32, Richard Marken <rsmarken@gmail.com> wrote:

> [From Rick Marken (2012.10.05.1530)]
>
> On Fri, Oct 5, 2012 at 12:16 PM, R K Moore <r.k.moore@dcs.shef.ac.uk> > wrote:
>
>> Thought you all might enjoy this TED talk by Daniel Wolpert ...
>>
http://www.ted.com/talks/daniel_wolpert_the_real_reason_for_brains.ht
>> ml
>>
>> ... lots of PCT-relevant arguments (including a role for
prediction).
>
> Well, I don't know if "enjoy" is quite the right word to describe my
> feeling about the talk (I couldn't make it all the way through,
> actually) but it was nice to see that the Very Serious People (VSP)
in
> robotics are no closer to understanding how things work than do the
> VSP in economics (VSP is Krugman's term for the people who are taken
> seriously in economics, and shouldn't be).
>
> The talk did have a lot of PCT relevant arguments in the sense that
> PCT has shown that all of the arguments are incorrect. I knew things
> were going south when Wolpert said that the brain is there to control
> movement. He's only off by ~180 degree. The brain, of course, is
there
> to control perceptions.
>
> It looks like the PCT secret is still safe;-)
>
> Best
>
> Rick
>
> --
> Richard S. Marken PhD
> rsmarken@gmail.com
> www.mindreadings.com

[From Rick Marken (2012.10.08.1800)]

Sigh!

RKM: Sorry to disappoint Warren, but when you say that you agree with Rick, do
you mean that that you agree ...
  - it's OK to voice an opinion on a subject (i.e. the content of Wolpert's
talk) without fully reviewing the material in question?

RM: I'm really sorry if my reply upset you Roger. I should have been
much more diplomatic. It's true that I didn't watch the entire TED
talk but I don't think it was necessary. You had posted the video
saying that there's lots of PCT-relevant arguments; but the whole
thing was about controlling actions (through Bayesian inference,
predictive control, action selection etc) and control of action is not
a part of PCT, nor has there been any data produced to show that it
needs to be a part of PCT. There may be situations where prediction is
necessary -- I think Powers has a model of eye movement that requires
some prediction -- but even when prediction is used it's part of a
closed loop control process that fixes things up after the
prediction-based movement is made.

But the big problem with Wolpert's talk, from my perspective, is that
he starts with the assumption that it is movement itself that is
controlled -- which is precisely wrong from the PCT perspective;
prediction or no prediction, what is controlled has to be a perceived
consequence of movement. Based on this assumption (control of output)
Wolpert concludes that prediction is always involved in the production
of movements.

PCT, on the other hand, starts with the assumption that it is the
perceptual consequences of movements -- not the movements themselves -
that are controlled. Models based on this assumption account for a
lot of movement data highly accurately without any need for the
incorporation of prediction. So PCT leads us to assume that prediction
is not involved in control until we are shown otherwise.

RKM: - it's acceptable to publicly question the veracity of someone's
scientific reputation without providing specific reasons?

RM: I don't believe I ever questioned the veracity of Wolpert's
scientific reputation. I was questioning the relevance of his work to
PCT. From a PCT perspective, Wolpert is on exactly the wrong track.
That doesn't mean he's a bad scientist; just that he doesn't seem to
have read any PCT and if he has he hasn't understood or been convinced
by what he's read. Heck, according to Einstein, the greatest
physicists of his time were wrong but that didn't mean that Einstein
was questioning their scientific reputations.

RKM: - it's productive to express satisfaction that a related field of study
seems not to have all the answers to a bunch of seriously hard problems that
face us all?

RM: Huh? To quote my friend Bob "No I don't feel so good when I see
the heartaches you embrace". I take no satisfaction when I see an
obviously very smart researcher headed down what I know is the wrong
path. But PCT has been rather harshly rejected by the "control of
movement" people (for what are clearly political rather than
scientific reasons; PCT has never been rejected by data) so it's a
little tiresome to have another "predictive control of output"
theorist trotted out as someone we could learn something from.

RKM: - it's correct to claim that "PCT has shown that all of the arguments are
incorrect" (i.e. Bayesian inference, predictive control, algorithms for
separating 'self' from 'other', action selection etc.) without citing the
relevant peer-reviewed evidence to back it up in each case?

RM: I didn't know that your post was supposed to be the start of such
a discussion. Predictive control has been a topic on CSGNet many times
but I would be happy to have that discussion again if you like. I can
try to point to relevant research and references once we get started.

RKM: I had hoped that people would see the potential overlaps and opportunities
to move our science and understanding forward collectively.

RM: I guess your hope (prediction) didn't pan out that well; that's
the problem with prediction; it's not nearly as reliable as control of
input;-) Perhaps it would have helped if you had used Bayesian
prediction. Then you could have included in your calculation the fact
that among those of us who do research on PCT the prior probability of
seeing overlap between a control of output model and PCT is extremely
low. But I am certainly willing to discuss this; if you can show me
some data that seems to demand Bayesian inference, predictive control
or action selection, then I could be convinced that there is overlap
between PCT and what Wolpert was talking about. So let's have the
discussion.

Best

Rick

···

On Mon, Oct 8, 2012 at 10:30 AM, R K Moore <r.k.moore@dcs.shef.ac.uk> wrote:
---
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

has anyone done a TED talk on PCT, if not, why not, the opportunities to do TED and TEDx talks are everywhere, do one and put it out there, I hear talk about the scientific community embracing PCT, but who cares if they embrace it first or not, with TED talks you reach the world first, the people least invested in years of education they’ve paid for to support their beliefs, and the most open to new ideas, I’d love to give one on it, but certainly not one of the people qualified to do so, if you want to change the world, here’s the opportunity that is but a phone call to one of your local Tedx talks away

Andrew Speaker

Lions For Change

3040 Peachtree Rd, Suite 312

Atlanta, Ga. 30305

404-913-3193

www.LionsForChange.com

“Go confidently in the direction of your dreams. Live the life you have imagined.” – Henry David Thoreau

···

On Oct 8, 2012, at 8:59 PM, Richard Marken wrote:

[From Rick Marken (2012.10.08.1800)]

On Mon, Oct 8, 2012 at 10:30 AM, R K Moore r.k.moore@dcs.shef.ac.uk wrote:

Sigh!

RKM: Sorry to disappoint Warren, but when you say that you agree with Rick, do
you mean that that you agree …

  • it’s OK to voice an opinion on a subject (i.e. the content of Wolpert’s
    talk) without fully reviewing the material in question?

RM: I’m really sorry if my reply upset you Roger. I should have been
much more diplomatic. It’s true that I didn’t watch the entire TED
talk but I don’t think it was necessary. You had posted the video
saying that there’s lots of PCT-relevant arguments; but the whole
thing was about controlling actions (through Bayesian inference,
predictive control, action selection etc) and control of action is not
a part of PCT, nor has there been any data produced to show that it
needs to be a part of PCT. There may be situations where prediction is
necessary – I think Powers has a model of eye movement that requires
some prediction – but even when prediction is used it’s part of a
closed loop control process that fixes things up after the
prediction-based movement is made.

But the big problem with Wolpert’s talk, from my perspective, is that
he starts with the assumption that it is movement itself that is
controlled – which is precisely wrong from the PCT perspective;
prediction or no prediction, what is controlled has to be a perceived
consequence of movement. Based on this assumption (control of output)
Wolpert concludes that prediction is always involved in the production
of movements.

PCT, on the other hand, starts with the assumption that it is the
perceptual consequences of movements – not the movements themselves -
that are controlled. Models based on this assumption account for a
lot of movement data highly accurately without any need for the
incorporation of prediction. So PCT leads us to assume that prediction
is not involved in control until we are shown otherwise.

RKM: - it’s acceptable to publicly question the veracity of someone’s
scientific reputation without providing specific reasons?

RM: I don’t believe I ever questioned the veracity of Wolpert’s
scientific reputation. I was questioning the relevance of his work to
PCT. From a PCT perspective, Wolpert is on exactly the wrong track.
That doesn’t mean he’s a bad scientist; just that he doesn’t seem to
have read any PCT and if he has he hasn’t understood or been convinced
by what he’s read. Heck, according to Einstein, the greatest
physicists of his time were wrong but that didn’t mean that Einstein
was questioning their scientific reputations.

RKM: - it’s productive to express satisfaction that a related field of study
seems not to have all the answers to a bunch of seriously hard problems that
face us all?

RM: Huh? To quote my friend Bob “No I don’t feel so good when I see
the heartaches you embrace”. I take no satisfaction when I see an
obviously very smart researcher headed down what I know is the wrong
path. But PCT has been rather harshly rejected by the “control of
movement” people (for what are clearly political rather than
scientific reasons; PCT has never been rejected by data) so it’s a
little tiresome to have another “predictive control of output”
theorist trotted out as someone we could learn something from.

RKM: - it’s correct to claim that “PCT has shown that all of the arguments are
incorrect” (i.e. Bayesian inference, predictive control, algorithms for
separating ‘self’ from ‘other’, action selection etc.) without citing the
relevant peer-reviewed evidence to back it up in each case?

RM: I didn’t know that your post was supposed to be the start of such
a discussion. Predictive control has been a topic on CSGNet many times
but I would be happy to have that discussion again if you like. I can
try to point to relevant research and references once we get started.

RKM: I had hoped that people would see the potential overlaps and opportunities
to move our science and understanding forward collectively.

RM: I guess your hope (prediction) didn’t pan out that well; that’s
the problem with prediction; it’s not nearly as reliable as control of
input;-) Perhaps it would have helped if you had used Bayesian
prediction. Then you could have included in your calculation the fact
that among those of us who do research on PCT the prior probability of
seeing overlap between a control of output model and PCT is extremely
low. But I am certainly willing to discuss this; if you can show me
some data that seems to demand Bayesian inference, predictive control
or action selection, then I could be convinced that there is overlap
between PCT and what Wolpert was talking about. So let’s have the
discussion.

Best

Rick

Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bill Powers (2012.10.09.0935 MDY)]

[From Rick Marken (2012.10.08.1800)]

> Sigh!
>
>RKM: Sorry to disappoint Warren, but when you say that you agree with Rick, do
> you mean that that you agree ...

I liked this reply to Roger Moore a lot better than the first one; in fact I admired this one for having a lot of substance. All that's really necessary is to go into some of the details. I agree with you that all the conventional approaches (that I know of) are built on a mistaken idea of how living control systems work. This rather subverts the intention of peer review because the peers generally believe in the same erroneous theories and take them for granted.

Here's another example for you. In neuroscience, most of the discussions of motor control you will read in this field assume that control is achieved by planning an action that will have a predicted outcome and then executing the action needed to achieve it. But the evidence on which this theory of control is based is tainted by the very same theory.

The evidence consists of finding activity in certain brain regions that starts before the control actions begin. The assumption is that this shows the predictions and planning of action taking place. Of course nobody has ever actually observed any predictions or any plans being created; what is observed is "activity." That's about as close as neuroscience ever gets to identifying the actual processes carried out by the neural activities. Nevertheless, this identification is regularly used in explanations in the peer-reviewed literature as if it were an established fact.

This preceding activity has a different explanation in PCT: it is assumed to indicate generation of reference signals. The generation of reference signals naturally precedes the actions that bring perceptions of the consequences to values that match the reference signals. An external observer could use observations of the reference signals as a means of predicting how the behavior is going to change, and as a consequence, how the perceptual signals are going to change. But that does not mean the system being observed is also making plans and predictions. Only the external observer is making any predictions. This proves that the human brain CAN make predictions, but it does not show that this is how motor control works.

Evidence for the PCT view also exists. E. Roy John, then at Johns Hopkins University, did experiments showing that a cat that is expecting some perceptual event to take place experiences perceptual signals in the same neurons of the brain that light up when the event actually does take place. So what we call the "imagination connection" is quite strongly implied. In other experiments this has been extended to the result of intended actions: the signal that occurs before the action happens shows up again in the same neurons when the consequences of that action do happen. This is like imagining a reference condition, then using it as a reference signal for the negative feedback control system that carries out the action and makes a matching perceptual signal representing the outcome of that action occur.

We still can't observe the details to corroborate our assumptions about what all those signals, before and after the action, mean. However, the timing and locations of the signals offer a clear alternative to the predict-plan-execute model, a far simpler and more plausible alternative. Of course we also have the behavioral evidence.

Best,

Bill P.

···

At 05:59 PM 10/8/2012 -0700, Richard Marken wrote:

On Mon, Oct 8, 2012 at 10:30 AM, R K Moore <r.k.moore@dcs.shef.ac.uk> wrote: