Grokbase Groups Perl ai April 2002
FAQ
Hi,

I'm wondering whether we want to create an ML:: namespace for
Machine Learning stuff. The reasons in favor might be:

1) "AI" is sort of a loaded term, and attracts its fair share
of wackos. For this reason, many researchers eschew it.

2) Many interesting things would fit better into "Machine
Learning" than they would into AI, just because they don't
really claim to be "intelligent", but they do learn from
observed data.

Reasons against might be:

1) "ML" isn't nearly as well-known a term as "AI", so casual
browsers of CPAN won't find things there as easily.

2) The AI:: namespace already exists and contains some
interesting things, and it would be a shame to split the
namespace without a good reason.

Comments?

I ask because I'm working on a rewrite of AI::Categorize, and
I'm wondering whether it might fit better under the ML::
namespace.

-Ken

Search Discussions

  • John Douglas Porter at Apr 21, 2002 at 12:19 pm

    Ken Williams wrote:
    I ask because I'm working on a rewrite of AI::Categorize, and
    I'm wondering whether it might fit better under the ML::
    namespace.
    My personal opionion: AI shouldn't even be a namespace, really.
    It should just be a category.

    ML is a good namespace name.

    --
    John Douglas Porter
  • Bharani Kumar at Apr 21, 2002 at 6:21 pm
    hi,
    Thats right. But ML is a category under AI in broad.
    isn't it? So can we split it that way?

    bharani
    --- Ken Williams wrote:
    Hi,

    I'm wondering whether we want to create an ML::
    namespace for
    Machine Learning stuff. The reasons in favor might
    be:

    1) "AI" is sort of a loaded term, and attracts its
    fair share
    of wackos. For this reason, many researchers eschew
    it.

    2) Many interesting things would fit better into
    "Machine
    Learning" than they would into AI, just because they
    don't
    really claim to be "intelligent", but they do learn
    from
    observed data.

    Reasons against might be:

    1) "ML" isn't nearly as well-known a term as "AI",
    so casual
    browsers of CPAN won't find things there as easily.

    2) The AI:: namespace already exists and contains
    some
    interesting things, and it would be a shame to split
    the
    namespace without a good reason.

    Comments?

    I ask because I'm working on a rewrite of
    AI::Categorize, and
    I'm wondering whether it might fit better under the
    ML::
    namespace.

    -Ken

    __________________________________________________
    Do You Yahoo!?
    Yahoo! Games - play chess, backgammon, pool and more
    http://games.yahoo.com/
  • Mark Rogaski at Apr 21, 2002 at 6:57 pm
    An entity claiming to be Ken Williams (kenw@ee.usyd.edu.au) wrote:
    :
    : I ask because I'm working on a rewrite of AI::Categorize, and
    : I'm wondering whether it might fit better under the ML::
    : namespace.
    :

    I think the ML:: namespace is an excellent idea, for the same reasons you
    mentioned. AI:: is vague at best and is better suited for killer space
    robots than agglomerative clustering.

    Mark

    --
    [] | "Girls in occupied countries always
    [] Mark 'Doc' Rogaski | get into trouble with soldiers," she
    [] wendigo@pobox.com | said, when I asked her what the Virgin
    [] | birth was. -- Florence King, CoaFSL
  • Mike Lambert at Apr 22, 2002 at 1:41 am

    Reasons against might be:

    1) "ML" isn't nearly as well-known a term as "AI", so casual
    browsers of CPAN won't find things there as easily.

    2) The AI:: namespace already exists and contains some
    interesting things, and it would be a shame to split the
    namespace without a good reason.
    3) There's already a programming language called ML, and this might
    increase confusion. Standard ML, and OCaml are two variants on it. While
    I'm not advocating the creation of a namespace for that language, since
    none exists for other languages, this programming language was the first
    thing that came to mind for a non-ai programmer like myself.

    Perhaps AI::ML::* might be a better namespace? If I we're looking for a
    module that adapts or changes, I would check the AI namespace first. While
    it might be a catch-all, it's a rather descriptive catch-all for these
    types of modules, and you can always subdivide the namespace if it's too
    broad for your tastes.


    Mike Lambert
    A non-ai-programmer lurking on an ai-programmer mailing list. :)
  • Tolkin, Steve at Apr 22, 2002 at 1:17 pm
    I completely agree. Both people who
    know AI and its history, and people who do not,
    are likely to look for, or at least notice, AI.
    ML will not do that.
    MachineLearning might -- but that is
    not the only possible candidate.
    For example to avoid the poor connotations of
    "Data Mining" the ACM named its special interest group
    SIGKDD, where KDD stands for Knowledge Discovery in Databases.

    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    I agree that adding subcategories below AI can be useful.
    But I too would expect AI::ML to be about the language
    ML.


    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments are by me,
    not Fidelity Investments, its subsidiaries or affiliates.
    -----Original Message-----
    From: Mike Lambert
    Sent: Sunday, April 21, 2002 9:42 PM
    To: Ken Williams
    Cc: Perl-AI
    Subject: Re: An ML:: namespace?

    Reasons against might be:

    1) "ML" isn't nearly as well-known a term as "AI", so casual
    browsers of CPAN won't find things there as easily.

    2) The AI:: namespace already exists and contains some
    interesting things, and it would be a shame to split the
    namespace without a good reason.
    3) There's already a programming language called ML, and this might
    increase confusion. Standard ML, and OCaml are two variants
    on it. While
    I'm not advocating the creation of a namespace for that
    language, since
    none exists for other languages, this programming language
    was the first
    thing that came to mind for a non-ai programmer like myself.

    Perhaps AI::ML::* might be a better namespace? If I we're
    looking for a
    module that adapts or changes, I would check the AI namespace
    first. While
    it might be a catch-all, it's a rather descriptive catch-all for these
    types of modules, and you can always subdivide the namespace
    if it's too
    broad for your tastes.


    Mike Lambert
    A non-ai-programmer lurking on an ai-programmer mailing list. :)
  • Paris Sinclair at Apr 22, 2002 at 4:03 pm

    On Mon, 22 Apr 2002, Tolkin, Steve wrote:

    I completely agree. Both people who
    know AI and its history, and people who do not,
    are likely to look for, or at least notice, AI.
    ML will not do that.
    MachineLearning might -- but that is
    not the only possible candidate.
    For example to avoid the poor connotations of
    "Data Mining" the ACM named its special interest group
    SIGKDD, where KDD stands for Knowledge Discovery in Databases.

    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    I agree that adding subcategories below AI can be useful.
    But I too would expect AI::ML to be about the language
    ML.


    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments are by me,
    not Fidelity Investments, its subsidiaries or affiliates.
    That was part of the original point, that because of the nonsense around
    the term AI, that something lower profile would be better.

    So when you say "but only AI has widespread name recognition," are you
    agreeing with the proposal, or disagreeing with it?

    --
    Paris
  • Tolkin, Steve at Apr 22, 2002 at 4:18 pm
    I think continuing to use the AI namespace
    is correct. I was agreeing with another poster
    who (I think) also had this position.
    The fact that AI had too much hype
    does not mean the name is the wrong one.

    (Aside: Years from now we'll probably still use
    the term "web services" even though
    that term also is associated with more
    hype than it can support.)

    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments are by me,
    not Fidelity Investments, its subsidiaries or affiliates.

    -----Original Message-----
    From: Paris Sinclair
    Sent: Monday, April 22, 2002 12:03 PM
    To: Tolkin, Steve
    Cc: Perl-AI
    Subject: RE: An ML:: namespace?

    On Mon, 22 Apr 2002, Tolkin, Steve wrote:

    I completely agree. Both people who
    know AI and its history, and people who do not,
    are likely to look for, or at least notice, AI.
    ML will not do that.
    MachineLearning might -- but that is
    not the only possible candidate.
    For example to avoid the poor connotations of
    "Data Mining" the ACM named its special interest group
    SIGKDD, where KDD stands for Knowledge Discovery in Databases.

    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    I agree that adding subcategories below AI can be useful.
    But I too would expect AI::ML to be about the language
    ML.


    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments
    are by me,
    not Fidelity Investments, its subsidiaries or affiliates.
    That was part of the original point, that because of the
    nonsense around
    the term AI, that something lower profile would be better.

    So when you say "but only AI has widespread name recognition," are you
    agreeing with the proposal, or disagreeing with it?

    --
    Paris
  • Jpnolan at Apr 30, 2002 at 6:31 pm

    --- Ken Williams wrote:
    1) "AI" is sort of a loaded term, and attracts its
    fair share of wackos. For this reason, many researchers
    eschew it.

    2) Many interesting things would fit better into
    "Machine Learning" than they would into AI, just because
    they don't really claim to be "intelligent", but they
    do learn from observed data.
    On Mon, 22 Apr 2002, Tolkin, Steve wrote:
    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    Speaking for myself, I just don't like the term
    "artificial intelligence," for the same reasons that other
    folks dislike it. So I would agree with the other posters,
    that another top-level category might be a good idea.

    However, I think some of the same complaints also apply
    to the term "ML::" or "MachineLearning::". What does
    "learning" really mean?

    In theory I like the idea of eliminating the term 'AI',
    however, I don't know what new name would be
    a good general replacement. :(

    My hunch is that there is an approprate term out there somewhere,
    but we haven't thought of it yet. The name has to say that
    the algorithms are "intelligent" but actually "not intelligent".
    Something like "AttemptedIntelligence::" or "AllegedlyIntelligent::".

    Or maybe "HAL-wanna-be::". :)


    --
    #-------------------------
    # John Nolan
    # jpnolan sonic net
    #-------------------------
  • BUDNEY, DANIEL L at Apr 30, 2002 at 7:12 pm

    the algorithms are "intelligent" but actually "not intelligent".
    Something like "AttemptedIntelligence::" or "AllegedlyIntelligent::".
    Isn't the Holy Grail to create (hence Artificial) something that is "really"
    Intelligent? Doesn't Artificial Intelligence describe the goal, and the
    things that go within that namespace contribute to that goal?
    I don't think anyone doing research on space navigation would look at
    "InterstellarTravel::" and expect that the stuff there actually resulted in
    someone going to another star.
    I do realize the media have sensationalized Artificial Intelligence to the
    point of pseudo-science, but people actually doing research in the field
    won't have such silly notions. (Will they?)
    If the goal is not Artificial Intelligence, then maybe you need a namespace
    describing what it is, instead of what it is working towards. Would
    "KnowledgeProcessing::" cover the things currently there? (It's beyond me
    how anyone will ever find anything under a new term like Knowledge
    Processing, but I'm doing my best to give the community options. I actually
    think "Knowledge-based Processing" would be even more appropriate, but
    that's way too long, and "KBP::" would be so ambiguous that you could give
    up hope of anyone ever finding anything there.)
    ----
    Where do forest rangers go to get away from it all?


    -----Original Message-----
    From: jpnolan
    Sent: Tuesday April 30, 2002 13:31
    To: perl-ai@perl.org
    Subject: Re: An ML:: namespace?

    --- Ken Williams wrote:
    1) "AI" is sort of a loaded term, and attracts its
    fair share of wackos. For this reason, many researchers
    eschew it.
    > >
    2) Many interesting things would fit better into
    "Machine Learning" than they would into AI, just because
    they don't really claim to be "intelligent", but they
    do learn from observed data.
    On Mon, 22 Apr 2002, Tolkin, Steve wrote:
    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    Speaking for myself, I just don't like the term
    "artificial intelligence," for the same reasons that other
    folks dislike it. So I would agree with the other posters,
    that another top-level category might be a good idea.

    However, I think some of the same complaints also apply
    to the term "ML::" or "MachineLearning::". What does
    "learning" really mean?

    In theory I like the idea of eliminating the term 'AI',
    however, I don't know what new name would be
    a good general replacement. :(

    My hunch is that there is an approprate term out there
    somewhere,
    but we haven't thought of it yet. The name has to say that
    the algorithms are "intelligent" but actually "not
    intelligent".
    Something like "AttemptedIntelligence::" or
    "AllegedlyIntelligent::".

    Or maybe "HAL-wanna-be::". :)


    --
    #-------------------------
    # John Nolan
    # jpnolan sonic net
    #-------------------------
  • Matt Youell at Apr 30, 2002 at 9:57 pm
    I don't have the full thread here in front of me right now, but I seem to
    recall the thread started with someone (kw?) looking to relocate
    AI::Categorize. Tasks like categorization, taxonomy, etc. seem to be less
    "AI" and more Knowledge Management (KM). Knowledge Management is an existing
    term that's common in Fortune 500 settings (vs. academia), and there are
    many active products exploring that market right now. It's a common enough
    term to be recognized, still broad, but it doesn't have the cultural baggage
    that AI does, imho.

    Just for comparison, "Artificial Intelligence" googles @ 1 M hits, while
    "Knowledge Management" racks up 756k.
    "Machine Learning" gets 286k

    KM:: ?


    ________________________________________
    matt youell
    http://www.youell.com/matt/
    "think different - just like everyone else."


    ----- Original Message -----
    From: "BUDNEY, DANIEL L" <DBUDN90@entergy.com>
    To: <perl-ai@perl.org>
    Sent: Tuesday, April 30, 2002 12:11 PM
    Subject: RE: An ML:: namespace?

    the algorithms are "intelligent" but actually "not intelligent".
    Something like "AttemptedIntelligence::" or "AllegedlyIntelligent::".
    Isn't the Holy Grail to create (hence Artificial) something that is "really"
    Intelligent? Doesn't Artificial Intelligence describe the goal, and the
    things that go within that namespace contribute to that goal?
    I don't think anyone doing research on space navigation would look at
    "InterstellarTravel::" and expect that the stuff there actually resulted in
    someone going to another star.
    I do realize the media have sensationalized Artificial Intelligence to the
    point of pseudo-science, but people actually doing research in the field
    won't have such silly notions. (Will they?)
    If the goal is not Artificial Intelligence, then maybe you need a namespace
    describing what it is, instead of what it is working towards. Would
    "KnowledgeProcessing::" cover the things currently there? (It's beyond me
    how anyone will ever find anything under a new term like Knowledge
    Processing, but I'm doing my best to give the community options. I actually
    think "Knowledge-based Processing" would be even more appropriate, but
    that's way too long, and "KBP::" would be so ambiguous that you could give
    up hope of anyone ever finding anything there.)
    ----
    Where do forest rangers go to get away from it all?


    -----Original Message-----
    From: jpnolan
    Sent: Tuesday April 30, 2002 13:31
    To: perl-ai@perl.org
    Subject: Re: An ML:: namespace?

    --- Ken Williams wrote:
    1) "AI" is sort of a loaded term, and attracts its
    fair share of wackos. For this reason, many researchers
    eschew it.

    2) Many interesting things would fit better into
    "Machine Learning" than they would into AI, just because
    they don't really claim to be "intelligent", but they
    do learn from observed data.
    On Mon, 22 Apr 2002, Tolkin, Steve wrote:
    There are many other terms that cover some or most
    of this area -- but only AI has
    widespread name recognition.

    Speaking for myself, I just don't like the term
    "artificial intelligence," for the same reasons that other
    folks dislike it. So I would agree with the other posters,
    that another top-level category might be a good idea.

    However, I think some of the same complaints also apply
    to the term "ML::" or "MachineLearning::". What does
    "learning" really mean?

    In theory I like the idea of eliminating the term 'AI',
    however, I don't know what new name would be
    a good general replacement. :(

    My hunch is that there is an approprate term out there
    somewhere,
    but we haven't thought of it yet. The name has to say that
    the algorithms are "intelligent" but actually "not
    intelligent".
    Something like "AttemptedIntelligence::" or
    "AllegedlyIntelligent::".

    Or maybe "HAL-wanna-be::". :)


    --
    #-------------------------
    # John Nolan
    # jpnolan sonic net
    #-------------------------
  • John Douglas Porter at May 1, 2002 at 4:38 am

    Matt Youell wrote:
    AI::Categorize. Tasks like categorization, taxonomy, etc. seem to be less
    "AI" and more Knowledge Management (KM).
    I dunno. That really depends on what you define as AI.
    Traditionally, the AI net is cast broadly enough to cover
    those things.

    Knowledge Management is an existing
    term that's common in Fortune 500 settings (vs. academia), and there are
    many active products exploring that market right now. It's a common enough
    term to be recognized, still broad,
    And this touches on the heart of the problem, as I've described in
    another email. It's not that "AI" carries a lot of baggage... So what?
    It's that the category is just TOO broad!



    I actually like "ML::", except for the fact that it would no doubt
    have other meanings to other people.
    (Someone said there is no precedent for top-level names for
    other languages. In fact, there is "C::", but admittedly
    that's a special case.)


    And I'm thinking that acronyms are probably suboptimal, in general.

    --
    John Douglas Porter
  • Matt Youell at May 1, 2002 at 6:53 am
    <snippage>
    And this touches on the heart of the problem, as I've described in
    another email. It's not that "AI" carries a lot of baggage... So what?
    It's that the category is just TOO broad!
    Yeah, I agree, it's too broad. But it's not just that. The Math:: category
    is arguably broader... and yet I don't cringe when I hear the word "Math".
    Ya know?

    <snippage>
    And I'm thinking that acronyms are probably suboptimal, in general.
    I'm with you there.



    ________________________________________
    matt youell
    http://www.youell.com/matt/
    Visit beautiful Einstein County, KY - "Where everyone's relative"
  • John Douglas Porter at May 1, 2002 at 2:27 pm

    Matt Youell wrote:
    it's too broad. But it's not just that. The Math:: category
    is arguably broader... and yet I don't cringe when I hear the word "Math".
    Well, I don't cringe when I hear "AI".
    So I'm not too sure your analogy works.

    --
    John Douglas Porter
  • Ianb at May 2, 2002 at 1:52 am

    On 30 Apr 02 at 11:53:48PM, Matt Youell wrote:
    ... and yet I don't cringe when I hear the word "Math".
    Ya know?
    I cringe when I hear the word "math", but only because where I come
    from (and any right-thinking person/AI entity would agree), it's
    called "maths".

    Regards,


    Ian Boreham

    (Humour not marked)
  • John Douglas Porter at May 1, 2002 at 4:12 am

    BUDNEY, DANIEL L wrote:
    Where do forest rangers go to get away from it all?
    To work.


    (Isn't it obvious?)


    --
    John Douglas Porter

    There are no Killing spinors on the brane.
  • John Douglas Porter at May 1, 2002 at 4:23 am

    BUDNEY, DANIEL L wrote:
    ...maybe you need a namespace
    describing what it is, instead of what it is working towards. Would
    "KnowledgeProcessing::" cover the things currently there?
    I think the essential problem is not that "AI" isn't meaningful enough,
    or carries too much baggage, but that the domain it's supposed to
    cover is just too damn broad.

    If (and we should be so fortunate) there were modules under AI:: for
    doing all the different kinds of things that people who do "AI" do,
    it would be the largest namespace on CPAN, by far.

    --
    John Douglas Porter
  • Tolkin, Steve at May 1, 2002 at 12:40 pm
    The same tar-filled brush that swiped AI will
    get KM too, becaue again the (hype+bs)/value
    ratio is very high. In fact the small pieces of
    KM that are worthwhile are based on a variety of
    techniques: mostly statistical natural language
    processing (This is not a mainstream part of linguistics
    for many years, because it does not pay any attention
    to even grammar. Rather it is IR (information retrieval)
    informed by AI, statistic, data mining, etc.

    Teh part of KM that is most like AI has the same
    probelms of it doesn't scale in size,
    knwoledge from one "domain" doesn't transfer
    easily to another domain, etc.


    The benefit of the AI namespace is that people
    will think to look there.
    No one who is not alreadyu aware of the acronym
    KM would look there. I too prefer to avoid
    acronyms except those that are already deeply
    entrenched in the culture -- AI is one such.

    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments are by me,
    not Fidelity Investments, its subsidiaries or affiliates.




    -----Original Message-----
    From: John Douglas Porter
    Sent: Wednesday, May 01, 2002 12:39 AM
    To: perl-ai@perl.org
    Subject: Re: An ML:: namespace?


    Matt Youell wrote:
    AI::Categorize. Tasks like categorization, taxonomy, etc.
    seem to be less
    "AI" and more Knowledge Management (KM).
    I dunno. That really depends on what you define as AI.
    Traditionally, the AI net is cast broadly enough to cover
    those things.

    Knowledge Management is an existing
    term that's common in Fortune 500 settings (vs. academia),
    and there are
    many active products exploring that market right now. It's
    a common enough
    term to be recognized, still broad,
    And this touches on the heart of the problem, as I've described in
    another email. It's not that "AI" carries a lot of
    baggage... So what?
    It's that the category is just TOO broad!



    I actually like "ML::", except for the fact that it would no doubt
    have other meanings to other people.
    (Someone said there is no precedent for top-level names for
    other languages. In fact, there is "C::", but admittedly
    that's a special case.)


    And I'm thinking that acronyms are probably suboptimal, in general.

    --
    John Douglas Porter

  • Don Blaheta at May 1, 2002 at 1:10 pm

    Quoth Tolkin, Steve:
    statistical natural language processing (This is not a mainstream part
    of linguistics for many years, because it does not pay any attention
    to even grammar. Rather it is IR (information retrieval) informed by
    AI, statistic, data mining, etc.
    I have to disagree. NLP (which is pretty much all statistical at this
    point in time) is quite a bit more than what you say. Also known as
    "computational linguistics", it covers nearly anything you can do with
    natural languages and computers. In particular, tagging and parsing are
    common tasks (and you better believe parsing pays attention to grammar! :).
    Other subfields include machine translation, voice recognition, dialogue
    processing, anaphor resolution, prosody analysis... the list goes on.
    Language modelling is one important thing that NLP/CL does, and that's a
    very linguistic thing: as our (computer) language models improve, we
    gain insight into the models floating around in our brains.

    Anyway, that's mostly tangential. Apropos the real topic of discussion,
    I'd just like to point out that despite all the stupid bs and hype in
    the popular media, most AI researchers still call themselves AI
    researchers. If they adopted some other name, the media would just get
    hold of that and do the same thing. Meanwhile, AI *does* have the
    advantage of being very recognisable, and in particular, *the thing
    people would think of* if that's what they were looking for. It's even
    recognisable in acronym form.

    --
    -=-Don Blaheta-=-=-dpb@cs.brown.edu-=-=-<http://www.cs.brown.edu/~dpb/>-=-
    "In AI, much of the I is in the beholder." --Manning & Schütze
  • Sean M. Burke at May 1, 2002 at 9:41 pm

    At 08:39 2002-05-01 -0400, Tolkin, Steve wrote:
    In fact the small pieces of KM that are worthwhile are based on a variety
    of techniques: mostly statistical natural language processing (This is not
    a mainstream part of linguistics for many years, because it does not pay
    any attention to even grammar.
    You allude to the concept of "mainstream linguistics".

    My experience in linguistics points quite strongly to one conclusion: it
    just doesn't matter what is supposedly "mainstream" linguistics, if there
    even is such a thing. For example, the fact that the formalists manage to
    crank out more grad students and articles than functionalists do, says
    nothing about who's right, since the approaches are incommensurate.
    In short, scholarship (much less academia) is not a democracy, any more
    than religion is.
  • Tolkin, Steve at May 1, 2002 at 2:41 pm
    You quoted me out of context. I agree with what you say
    about NLP and computational linguistics.
    My point about KM (Knowledge Management) was that
    most products in this area use very little of this.

    Anyway, thanks for agreeing that keeping "AI" is perhaps
    the most appropriate name for the namespace.

    Hopefully helpfully yours,
    Steve
    --
    Steven Tolkin steve.tolkin@fmr.com 617-563-0516
    Fidelity Investments 82 Devonshire St. V8D Boston MA 02109
    There is nothing so practical as a good theory. Comments are by me,
    not Fidelity Investments, its subsidiaries or affiliates.

    -----Original Message-----
    From: Don Blaheta
    Sent: Wednesday, May 01, 2002 9:10 AM
    To: perl-ai@perl.org
    Subject: Re: An ML:: namespace?


    Quoth Tolkin, Steve:
    statistical natural language processing (This is not a
    mainstream part
    of linguistics for many years, because it does not pay any attention
    to even grammar. Rather it is IR (information retrieval)
    informed by
    AI, statistic, data mining, etc.
    I have to disagree. NLP (which is pretty much all statistical at this
    point in time) is quite a bit more than what you say. Also known as
    "computational linguistics", it covers nearly anything you can do with
    natural languages and computers. In particular, tagging and
    parsing are
    common tasks (and you better believe parsing pays attention
    to grammar! :).
    Other subfields include machine translation, voice
    recognition, dialogue
    processing, anaphor resolution, prosody analysis... the list goes on.
    Language modelling is one important thing that NLP/CL does,
    and that's a
    very linguistic thing: as our (computer) language models improve, we
    gain insight into the models floating around in our brains.

    Anyway, that's mostly tangential. Apropos the real topic of
    discussion,
    I'd just like to point out that despite all the stupid bs and hype in
    the popular media, most AI researchers still call themselves AI
    researchers. If they adopted some other name, the media
    would just get
    hold of that and do the same thing. Meanwhile, AI *does* have the
    advantage of being very recognisable, and in particular, *the thing
    people would think of* if that's what they were looking for.
    It's even
    recognisable in acronym form.

    --
    -=-Don
    Blaheta-=-=-dpb@cs.brown.edu-=-=-<http://www.cs.brown.edu/~dpb/>-=-
    "In AI, much of the I is in the beholder." --Manning & Schütze
  • Matt Youell at May 1, 2002 at 4:25 pm

    You quoted me out of context. I agree with what you say
    about NLP and computational linguistics.
    My point about KM (Knowledge Management) was that
    most products in this area use very little of this.
    I left my opinion unclear, and I'm just realizing it after reading this
    morning's batch of posts. I didn't mean to suggest AI:: go away. I meant for
    KM:: (or whatever - your pet name here) to be an additional category
    parallel to it, for libs that the authors feel don't quite fit into AI for
    one reason or another. I get the vibe that this still would not be popular,
    though. Just a thought.

    Anyway, thanks for agreeing that keeping "AI" is perhaps
    the most appropriate name for the namespace.
    There does seem to be a consensus for this. I suppose if It Ain't Broke...



    ________________________________________
    matt youell
    http://www.youell.com/matt/
    "It's hard to get people to drink the Kool-Aid."
  • Lee Goddard at May 2, 2002 at 1:58 pm

    At 23:53 30/04/2002 -0700, Matt Youell wrote:
    I don't cringe when I hear the word "Math".
    I do: I'm English; we say 'maths'.

    Sorry
    lee
  • Lee Goddard at May 2, 2002 at 1:58 pm

    At 00:39 01/05/2002 -0400, John Douglas Porter wrote:
    (Someone said there is no precedent for top-level names for
    other languages. In fact, there is "C::", but admittedly
    that's a special case.)
    There is also a top-level named Language, which is
    apparently the recommended place for ... languages.

    lee
  • Lee Goddard at May 2, 2002 at 1:58 pm

    At 23:53 30/04/2002 -0700, Matt Youell wrote:
    I don't cringe when I hear the word "Math".
    I do: I'm English; we say 'maths'.

    Sorry
    lee
  • Lee Goddard at May 2, 2002 at 1:58 pm

    At 00:39 01/05/2002 -0400, John Douglas Porter wrote:
    (Someone said there is no precedent for top-level names for
    other languages. In fact, there is "C::", but admittedly
    that's a special case.)
    There is also a top-level named Language, which is
    apparently the recommended place for ... languages.

    lee
  • Lee Goddard at May 2, 2002 at 1:58 pm

    At 23:53 30/04/2002 -0700, Matt Youell wrote:
    I don't cringe when I hear the word "Math".
    I do: I'm English; we say 'maths'.

    Sorry
    lee
  • Matt Youell at May 2, 2002 at 4:28 pm

    At 23:53 30/04/2002 -0700, Matt Youell wrote:
    I don't cringe when I hear the word "Math".
    I do: I'm English; we say 'maths'.
    There seems to be a bit of that going 'round. We could create a new category
    called Maths:: for those speaking the Queen's English... =^) But God help
    us if there is ever an Aluminum:: category!

    (Btw, I hearby propose Rithmetic:: for code from the rural US)


    ________________________________________
    matt youell
    http://www.youell.com/matt/
    "It's hard to get people to drink the Kool-Aid."
  • Mark Rogaski at May 2, 2002 at 6:08 pm
    An entity claiming to be Matt Youell (matt@youell.com) wrote:
    :
    : There seems to be a bit of that going 'round. We could create a new category
    : called Maths:: for those speaking the Queen's English... =^) But God help
    : us if there is ever an Aluminum:: category!
    :
    : (Btw, I hearby propose Rithmetic:: for code from the rural US)
    :

    Oh, let's go whole hog and debate the Hoagie:: namespace.

    Mark

    --
    [] | Use what talents you possess: the woods
    [] Mark 'Doc' Rogaski | would be very silent if no birds sang
    [] wendigo@pobox.com | except those that sang best.
    [] | -- Henry Van Dyke
  • Lee Goddard at May 6, 2002 at 3:00 pm

    At 09:28 02/05/2002 -0700, Matt Youell wrote:
    (Btw, I hearby propose Rithmetic:: for code from the rural US)
    And it'd comply nicely with the existing Three Rs namespace.





    Lee Goddard
    perl -e "while(1){print rand>0.5 ? chr 47 : chr 92}"

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupai @
categoriesperl
postedApr 21, '02 at 6:29a
activeMay 6, '02 at 3:00p
posts30
users14
websiteperl.org

People

Translate

site design / logo © 2021 Grokbase