Grokbase Groups Perl ai August 2001
FAQ
I've been reading a bit on molecular biology and bioinformatics, and I
really felt resonance with Danny Hillis's definition of AI as anything
we're using computers for but don't understand yet. I'm amazed at how
much noisy information is being generated by the various sequencing
projects, and how much of a black art the interpretation of the
results are.

I understand that there are a lot of people in AI trying to make sense
of the dirty data using classification and so on. Are any of those
people on this list? I'd love to hear what you're working on.

Nat

Search Discussions

  • Simon Cozens at Aug 12, 2001 at 7:18 pm

    On Sat, Aug 11, 2001 at 03:02:25PM -0600, Nathan Torkington wrote:
    I've been reading a bit on molecular biology and bioinformatics, and I
    really felt resonance with Danny Hillis's definition of AI as anything
    we're using computers for but don't understand yet.
    I think of AI as the study of programming situations where either don't
    know what you want, or don't know how to get it.
    - Sean Burke
    I understand that there are a lot of people in AI trying to make sense
    of the dirty data using classification and so on. Are any of those
    people on this list? I'd love to hear what you're working on.
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.

    Simon
    --
    He who makes a beast of himself gets rid of the pain of being a man.
    -H.S. Thompson
  • Nathan Torkington at Aug 12, 2001 at 10:21 pm

    Simon Cozens writes:
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.
    Thanks, but I already know a few. I was more interested in whether
    there were AI people doing work in the field. If the one response
    I've received is anything to go by, I guess not :-)

    Nat
  • Lee Goddard at Aug 12, 2001 at 11:26 pm

    Simon Cozens writes:
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.
    Thanks, but I already know a few. I was more interested in whether
    there were AI people doing work in the field. If the one response
    I've received is anything to go by, I guess not :-)
    I know uni depts where such work takes place, but not in Perl.
    Might you consider how to port a more trad AI lang to perl?
  • Stas Bekman at Aug 13, 2001 at 8:12 am

    On Sun, 12 Aug 2001, Nathan Torkington wrote:

    Simon Cozens writes:
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.
    Thanks, but I already know a few. I was more interested in whether
    there were AI people doing work in the field. If the one response
    I've received is anything to go by, I guess not :-)
    Nat, how do you know that if none of them is hanging out here, that they
    don't exist?

    google comes up with 15,200 hits
    http://www.google.com/search?q=bioinformatics+perl
    you may want to do some browsing...


    _____________________________________________________________________
    Stas Bekman JAm_pH -- Just Another mod_perl Hacker
    http://stason.org/ mod_perl Guide http://perl.apache.org/guide
    mailto:stas@stason.org http://localhost/ http://eXtropia.com/
    http://singlesheaven.com http://perl.apache.org http://perlmonth.com/
  • Lee Jones at Aug 14, 2001 at 9:25 am
    Nat:

    Personally my background is in CS and genetics -- within CS my
    concentration was in machine learning and algorithms. I've been working
    in bioinformatics for 5 or 6 years now and I'd have to say as the data
    sets become larger and larger they lend themselves more and more to a
    machine learning AI approach to analysis.

    Initially algorithms based on searching large spaces -- GAs, beam searches
    etc, but AI classification schemes are used quite a bit in
    specific analysis. I did a some work using KBANN (knowledge based
    artificial neural nets) for exon prediction, markov models, baysian
    classification networks etc for many different sub problems in biology and
    bioinformatics. Text classification schemes are used somewhat naively in
    trying to do automatic annotation also -- but I see a lot of room for
    improvement there.

    If you are interested or would like to know more I can talk or try to
    point you in some relevant directions.

    -lee

    On Sun, 12 Aug 2001, Nathan Torkington bestowed the following wisdom:
    Simon Cozens writes:
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.
    Thanks, but I already know a few. I was more interested in whether
    there were AI people doing work in the field. If the one response
    I've received is anything to go by, I guess not :-)

    Nat
  • Lenzo at Aug 16, 2001 at 11:21 am
    Well, part of the problem is the terminology. AI is dead. AI overpromised
    and underdelivered every time. All human understanding in 5 years,
    speech, vision, planning, everything -- it's 5 years off. 5 sets of
    5 years later, AI is still playing with toy problems.

    The terms today are "Information theory" and "Machine learning." Anyone
    on the bleeding edge of the technology looks at the term "AI" and just
    laughs a little inside. Information theory and machine learning
    mean nothing as labels, but the terms have grown out the of the ashes
    of "AI", and imply a lot more discipline and mathematical rigor.

    kevin "firebomb" lenzo

    On Sun, Aug 12, 2001 at 04:20:55PM -0600, Nathan Torkington wrote:
    Simon Cozens writes:
    I can put you in touch with a couple of bioinformatics/Perl people, but
    I don't know how much or if that would help.
    Thanks, but I already know a few. I was more interested in whether
    there were AI people doing work in the field. If the one response
    I've received is anything to go by, I guess not :-)

    Nat
  • Simon Cozens at Aug 17, 2001 at 10:42 am

    On Thu, Aug 16, 2001 at 07:15:56AM -0400, lenzo@cs.cmu.edu wrote:
    Well, part of the problem is the terminology. AI is dead.
    Shit. Guess we'll have to close the list down. :)
    The terms today are "Information theory" and "Machine learning."
    And "information retrieval". See some of Dan Brian's modules. The trick,
    though, is to lose the word "intelligence"; machines can't be
    intelligent, despite what Hofstadter thinks. It's all about doing things
    in a "human-like" way: how we look through data to find what we want,
    how we decide what's important and what isn't, and so on. We don't have
    to do that "intelligently", it's just that the tasks supposedly require
    intelligence. If you consider the average human who performs such tasks,
    it's suprising we suppose this.

    --
    Ermine? NO thanks. I take MINE black.
    - Henry Braun is Oxford Zippy
  • Don Blaheta at Aug 17, 2001 at 1:08 pm

    Quoth lenzo@cs.cmu.edu:
    The terms today are "Information theory" and "Machine learning." Anyone
    on the bleeding edge of the technology looks at the term "AI" and just
    laughs a little inside. Information theory and machine learning
    mean nothing as labels, but the terms have grown out the of the ashes
    of "AI", and imply a lot more discipline and mathematical rigor.
    Bullshit. AI is a broad field that encompasses things like machine
    learning and info theory, as well as planning, reasoning, natural
    language processing; all of these things have in common their attempt to
    give computers some aspect of what we call intelligence. Nobody is
    claiming (anymore) that we're "this close" to passing the Turing test,
    but we are making progress, each in our own way. And to suggest that
    people working in AI lack mathematical rigour, well, that's just silly.

    --
    -=-Don Blaheta-=-=-dpb@cs.brown.edu-=-=-<http://www.cs.brown.edu/~dpb/>-=-
    "[Pagers] retain the upper hand over mobile phones, thanks to fears the
    latter may interfere with delicate hospital equipment. At least that's
    what your doctors will tell you if they trade in their pager for a new
    putter or four iron." --BBC
  • Lenzo at Aug 18, 2001 at 12:35 pm
    My question is, why are you on the AI mailing list?
    Because I _am_ one of those AI guys. It's not the grand vision of AI
    that's broken, it's the public perception of the term. When someone
    says they do "AI" it conjures a whole set of concepts quite different
    from the other terms that have been deliberately chosen to avoid them.
    "Expert systems" were "AI", but their salesmen chose a different
    nomenclature in order to reach different funding source, particularly
    in industry. Today "Data mining" may be a buzzword of choice. In
    universities departments, papers, and in grant proposals, you will
    certainly find the shift, with "AI" virtually disappearing and people
    start talking about the problems they are working on, rather than
    "strong AI" and the simulation of intelligence. "Bioinformatics" is
    a clear example of the naming phenomenon: it's new, it's sexy,
    and hey it's informatics. It's more contentful and to the point.

    kevin
  • David M Guzak at Aug 23, 2001 at 12:12 am
    Kevin,

    Is this why you chose the signoff "firebomb" ? <G>
    From my understanding your first post was saying there was a problem with
    terminology. Then you go on to say "AI is dead" and over-promised and in
    your next post proudly proclaim to be one of those AI guys. I know this is
    quoting you out of context, but I'm having a hard time understanding your
    intent. It is good to hear your one of us.

    Sure there is a lot of hype lumped into the term AI, but to say
    bioinformatics is sexy just because they have not been associated with
    tremendous failures yet seems a little naïve. When used cars are sold as
    "pre-owned vehicles", it may make the tiniest difference to the salesmen
    bean counters, but it's all the same to the car mechanics. So what is it
    your trying to say to your fellow mechanics?

    And try not to use too many big words, you know how we mechanics can be.
    Lets talk more about the nuts and bolts and less about generalizations.

    -Dave



    -----Original Message-----
    From: lenzo@cs.cmu.edu
    Sent: Saturday, August 18, 2001 5:30 AM
    My question is, why are you on the AI mailing list?
    Because I _am_ one of those AI guys. It's not the grand vision of AI
    that's broken, it's the public perception of the term. When someone
    says they do "AI" it conjures a whole set of concepts quite different
    from the other terms that have been deliberately chosen to avoid them.
    "Expert systems" were "AI", but their salesmen chose a different
    nomenclature in order to reach different funding source, particularly
    in industry. Today "Data mining" may be a buzzword of choice. In
    universities departments, papers, and in grant proposals, you will
    certainly find the shift, with "AI" virtually disappearing and people
    start talking about the problems they are working on, rather than
    "strong AI" and the simulation of intelligence. "Bioinformatics" is
    a clear example of the naming phenomenon: it's new, it's sexy,
    and hey it's informatics. It's more contentful and to the point.

    kevin
  • Integer at Aug 18, 2001 at 5:10 am

    On Thu, Aug 16, 2001 at 07:15:56AM -0400, lenzo@cs.cmu.edu wrote:
    Well, part of the problem is the terminology. AI is dead.
    Shit. Guess we'll have to close the list down. :)
    The terms today are "Information theory" and "Machine learning."
    And "information retrieval". See some of Dan Brian's modules. The trick,
    though, is to lose the word "intelligence"; machines can't be
    intelligent, despite what Hofstadter thinks. It's all about doing things
    in a "human-like" way: how we look through data to find what we want,
    how we decide what's important and what isn't, and so on. We don't have
    to do that "intelligently", it's just that the tasks supposedly require
    intelligence. If you consider the average human who performs such tasks,
    it's suprising we suppose this.

    --
    Ermine? NO thanks. I take MINE black.
    - Henry Braun is Oxford Zippy



    machines can be intelligent. [as most humans]

    machines cannot be conscious. [as most humans]




    some did it out of self-doubt.
    you will do it because you like to experiment
    and see things with `your` `own` eyes.


    nn.






    /_/
    /
    \ \/ i should like to be a human plant
    \/ __
    __/
    i will shed leaves in the shade
    \_\ because i like stepping on bugs



    *--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--
    Netochka Nezvanova nezvanova@eusocial.com
    http://www.eusocial.com
    http://www.biohakc.com
    http://www.ggttctttat.com/!
    n r . 5 !!! http://steim.nl/leaves/petalz
    *--*--*--*--*--*--*--*--*--*--*--*--*-- --*--*--*--*--*--*--

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupai @
categoriesperl
postedAug 11, '01 at 9:03p
activeAug 23, '01 at 12:12a
posts12
users9
websiteperl.org

People

Translate

site design / logo © 2021 Grokbase