Re: Does robotic engineering have a place in strong AI research?


Scott T. Jensen wrote:
> Years ago in this newsgroup, there was a discussion about the seven main
> disciplines (a.k.a. academic fields) that make up strong AI research. They
> are as followed:
>
> Philosophy
> Mathematics
> Computational Neuroscience
> Electronic engineering (a.k.a. electrical and computer engineering)
> Informational Retrieval
> Cognitive Psychology
> Computer Science (emphasis on artificial intelligence)
>
> What I'm wondering is if robotic engineering has a place in the advance of
> AI research. If not, why not? If so, could it claim a spot from one of
> those above or would it rank eighth after them? If it can claim another's
> spot, which one could it claim and on what basis could it claim to be more
> important for strong AI research?


Forgetting about the issue of strong vs weak AI, and just speaking of
AI in general .... Rodney Brooks is famous for saying "simulations are
doomed to always succeed". Look up some of his discussions regards
situated cognition and embodied intelligence.

IOW, a simulation in a box is just that, a simulation in a box - and
you need to get the thing to actually work in the "real-world" to show
you have a something that actually works in the real-world. Saying so
isn't doing so. To give an analogy from physics, if all you ever did
was to play with equations, you might never discover that friction
exists. To give an analogy from engineering, in order to make the
models we use mathematically tractable, we have to make many
simplifications regarding how those models relate to the real-world
things they are modeling. Then, when we translate the models into
real-world devices, we have to figure out how to deal with the
differences between how the simplified models theoretically work, and
how physical devices actually work in the real-world. Saying so isn't
doing so.

.

Relevant Pages

  • Re: Does robotic engineering have a place in strong AI research?
    ... Electronic engineering ... Cognitive Psychology ... you need to get the thing to actually work in the "real-world" to show you have a something that actually works in the real-world. ...
    (comp.ai.philosophy)
  • Re: [OT] One for the iPhone/iTouch users
    ... He needs to remember there's a difference between formal Computer ... Science and 99% of real-world application development. ... I don't disagreed with what you're saying but there are shades of grey ...
    (uk.rec.motorcycles)
  • Re: Whose Fish - OO solution
    ... I was thinking of a way to make this problem more real-world instead ... The problem was fine as it was IMO, ... my understanding is that simulation tools are becoming ... will be entered by domain experts, not programmers for the most part. ...
    (comp.object)
  • Using CEP to determine random value
    ... I am creating a simple simulation using MS Excel. ... device provides an output value conforming to a CEP of 5 (e.g. 50% of ... getting lost as how to transform it into the CEP above. ... This is for a real-world job, ...
    (sci.stat.math)
  • Re: Pitch vs. trim in flight phases
    ... Steve Foley writes: ... including many real-world pilots--at least those with open ... minds. ... There is much to be gained from simulation. ...
    (rec.aviation.piloting)