Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article
From: Wolf Kirchmeir (wwolfkir_at_sympatico.ca)
Date: 07/30/04
- Next message: Sergio Navega: "Re: General Problem Solving and Consciousness"
- Previous message: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- In reply to: Eray Ozkural exa: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Next in thread: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Reply: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Reply: David Longley: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Messages sorted by: [ date ] [ thread ]
Date: Fri, 30 Jul 2004 09:27:38 -0400
Eray Ozkural exa wrote:
...snip..
> You really should look into the works of Ray Solomonoff for examples
> of that (and of course other machine learning algorithms at idsia.ch)
> if you are curious. It's mind-widening.
I'll check.
> In my opinion, classical approaches to machine learning are not
> general enough, for instance they do not address transfer of learning
> problem.
What do you mean by this?
If you mean, "How does learning in domain A become available for use in
domain B?", my experience as a teacher may provide a clue. If you mean
something else, the following may provide some amusement.
The traditional advice on how to study for an exam assumed that
"transfer of learning" was the key. You know, teachers told you to go
over your notes, to highlight passages of text, make outlines, memorise
formulas, and so on. Then you'd know the stuff you needed to use to
write the exam successfully. That was the advice I received, anyhow, and
the advice I passed on for the first 10 years or so of my teaching career.
Trouble is, on the exam you're not asked to go over your notes,
highlight passages of text, make outlines, memorise formulas, etc.
You're asked to solve problems, write essays, label diagrams, and so on.
And many students compplained that they had "studied" for hours before
the exam, yet couldn't succeed.
So eventually I told my students that to study for an exam they should
write essays, solve problems, label diagrams, and so on. That is,
"Practice what you will have to do on the exam." It seemed to me that
relying on "transfer of learning" was, well, unreliable.
Then there was the perennial complaint by my colleagues that their
students couldn't write essays, sometimes accompanied by a point-blank
"Do you actually teach that in your English classes?" I assured them I
taught my students how to summarise and analyse stories, etc. But
apparently these skills don't apply to writing a science paper. :-)
I also noticed that my students became very proficient at spelling
tests, then misspelled words at a wonderful rate in their essays...
IOW, "writing an essay" and "spelling" are context-bound like any other
skill set, and "transfer of learning" is difficult, to say the least.
If by "transfer of learning" you mean what happens when a
teacher-student relationship works, well, the only clue I have is that
the student learns what (s)he does. For example, if a student memorises
and recalls definitions of terms, then the student can memorise and
recall definitions of terms. But that does not mean that the student can
understand a text using those terms. Most of them can't even rephrase a
definiton "in your own words."
HTH
- Next message: Sergio Navega: "Re: General Problem Solving and Consciousness"
- Previous message: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- In reply to: Eray Ozkural exa: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Next in thread: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Reply: Sergio Navega: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Reply: David Longley: "Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|