Lojban In General

Posted by baisong on Tue 02 of Feb, 2010 23:55 GMT posts: 99

• sorry if I'm double-posting*

As for where 'getting computers understanding predicate logic' is, I'm
not sure what to say, but I always thought that first-order predicate
logic was the foundation of declarative languages like prolog.

Anyhow, as I understand the state of Machine Translation,
everyone is investigating hybrid models of various dynamic
programming algorithms to raise BLEU scores-- everyone's invested in
the statistical (i.e. NOT interlingual) approach, and trying to figure
out how to maximize performance. I think that there would be real
advantages to using lojban as an interlingual medium, instead of
essentially trying to imitate humans through machine learning, but
when you have huge corpora of bilingual (or better) natural language
data and virtually no bilingual corpora with lojban, it's just
infeasible. There were a few attempts to use Esperanto for that
purpose a while back (even before this paper), and no one seems to
cite them, except as failures.

Here's a really good up-to-date intro to MT:
http://docs.google.com/Doc?docid=0AYZKIeNnTBe2ZGd4azRrZm1fNTI2ZnpnYmRrZ2g

And a powerpoint version with lots of diagrams:
http://people.csail.mit.edu/people/koehn/publications/tutorial2003.pdf

Here's the origin of those two documents, if you're interested in more:
http://www.statmt.org/

Well, the sense in which computers cannot "understand" FOPL is limited, having to do with automatic proof recovery and the like (always running afoul of undecidability — the cutoff point in reduction chains). As far back (a least) as 1961, when I was discussing Loglan as an interlingua at the ALDP group at RAND, the use of FOPL was considered a positive advantage (the unique parsing as always — never mind that Loglan didn't have it then), though not enough to go forward with this line of development. The problem with FOPL is unsolvable, of course, though some types of meta procedures can push the practical limits back a long way (the meta proof that a particular line of reductions will never involve a conflict, for example).

From: "Jon "Top Hat" Jones" <eyeonus@gmail.com>
To: lojban-list@lojban.org
Sent: Tue, December 22, 2009 1:30:12 AM
Subject: lojban interlingua translation and first-order logic

I recently came across this paper, which discusses various methods of machine translation methods. In it it is mentioned that computers are not able to understand first-order (i.e. predicate) logic. Since the paper is nearly 2 decades old, I was wondering if anyone here knows what progress there has been in making it understandable by computers.

--
mu'o mi'e .aionys.

.i.a'o.e'e ko klama le bende pe denpa bu

Posted by efidetum on Wed 03 of Feb, 2010 00:12 GMT posts: 9

> what progress there
Even progress will not prove 2=3 (for instance), hopefully. Nor will it
revive Gödel <http://de.wikipedia.org/wiki/Kurt_G%C3%B6del>.

2009/12/21 Jon "Top Hat" Jones <eyeonus@gmail.com>:
> I recently came across this paper, which discusses various methods of
> machine translation methods. In it it is mentioned that computers are not
> able to understand first-order (i.e. predicate) logic. Since the paper is
> nearly 2 decades old, I was wondering if anyone here knows what progress
> there has been in making it understandable by computers.
>
>
> --
> mu'o mi'e .aionys.
>
> .i.a'o.e'e ko klama le bende pe denpa bu
>
>

Posted by efidetum on Wed 03 of Feb, 2010 00:13 GMT posts: 9

> first-order predicate logic was the foundation of declarative languages
In a political sense (please forgive this dirty word)? Anyway, there are
other (more modern) forms of logic.

> of declarative languages like prolog
Which is the utmost a computer is capable of!

> everyone is investigating hybrid models of various dynamic programming
algorithms
Not in particular.

> instead of essentially trying to imitate humans through machine learning
I am not supposed to lend out my abacus, am I?

2009/12/21 Oren <get.oren@gmail.com>

> *sorry if I'm double-posting*
>
> As for where 'getting computers understanding predicate logic' is, I'm
> not sure what to say, but I always thought that first-order predicate
> logic was the foundation of declarative languages like prolog.
>
> Anyhow, as I understand the state of Machine Translation,
> everyone is investigating hybrid models of various dynamic
> programming algorithms to raise BLEU scores-- everyone's invested in
> the statistical (i.e. NOT interlingual) approach, and trying to figure
> out how to maximize performance. I think that there would be real
> advantages to using lojban as an interlingual medium, instead of
> essentially trying to imitate humans through machine learning, but
> when you have huge corpora of bilingual (or better) natural language
> data and virtually no bilingual corpora with lojban, it's just
> infeasible. There were a few attempts to use Esperanto for that
> purpose a while back (even before this paper), and no one seems to
> cite them, except as failures.
>
> Here's a really good up-to-date intro to MT:
> http://docs.google.com/Doc?docid=0AYZKIeNnTBe2ZGd4azRrZm1fNTI2ZnpnYmRrZ2g
>
> And a powerpoint version with lots of diagrams:
> http://people.csail.mit.edu/people/koehn/publications/tutorial2003.pdf
>
> Here's the origin of those two documents, if you're interested in more:
> http://www.statmt.org/
>

I had a brief research position at university of Rochester (NY state)
under Dan Gildea, where they have a close-knit AI, computational
linguistics and neurological programs. I worked on a CYK decoder for
Chinese English machine translation, and the first several weeks of my
research were basically coming to terms with the history Machine
translation outlined in that paper; the rise of computational
linguistics in the 50s and 60s and then the adaptation to statistical
models in the 80s and 90s.

As for where 'getting computers understanding predicate logic' is, I'm
not sure what to say, but I always thought that first-order predicate
logic was the foundation of declarative languages like prolog.

Anyhow, the state of the program at U of R (which I'm told is pretty
advanced) is investigating hybrid models of various dynamic
programming algorithms to raise BLEU scores-- everyone's invested in
the statistical (i.e. NOT interlingual) approach, and trying to figure
out how to maximize performance. I think that there would be real
advantages to using lojban as an interlingual medium, instead of
essentially trying to imitate humans through machine learning, but
when you have huge corpora of bilingual (or better) natural language
data and virtually no bilingual corpora with lojban, it's just
infeasible. There were a few attempts to use Esperanto for that
purpose a while back (even before this paper), and no one seems to
cite them, except as failures.

Here's a really good up-to-date intro to MT:
http://docs.google.com/Doc?docid=0AYZKIeNnTBe2ZGd4azRrZm1fNTI2ZnpnYmRrZ2g

And a powerpoint version with lots of diagrams:
http://people.csail.mit.edu/people/koehn/publications/tutorial2003.pdf

mu'o mi'e .ku'us.

On 2009-12-21, Jon "Top Hat" Jones <eyeonus@gmail.com> wrote:
> I recently came across
> this<http://www.lojban.org/files/why-lojban/mactrans.txt>paper, which
> discusses various methods of machine translation methods. In it
> it is mentioned that computers are not able to understand first-order (i.e.
> predicate) logic. Since the paper is nearly 2 decades old, I was wondering
> if anyone here knows what progress there has been in making it
> understandable by computers.
>
>
> --
> mu'o mi'e .aionys.
>
> .i.a'o.e'e ko klama le bende pe denpa bu
>

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

Wow, sorry for those typos! (typoes?)

Here's the origin of those two documents, if you're interested in more:
http://www.statmt.org/

On 2009-12-21, 白松 Oren <baisong@gvbchina.org.cn> wrote:
> I had a brief research position at university of Rochester (NY state)
> under Dan Gildea, where they have a close-knit AI, computational
> linguistics and neurological programs. I worked on a CYK decoder for
> Chinese English machine translation, and the first several weeks of my
> research were basically coming to terms with the history Machine
> translation outlined in that paper; the rise of computational
> linguistics in the 50s and 60s and then the adaptation to statistical
> models in the 80s and 90s.
>
> As for where 'getting computers understanding predicate logic' is, I'm
> not sure what to say, but I always thought that first-order predicate
> logic was the foundation of declarative languages like prolog.
>
> Anyhow, the state of the program at U of R (which I'm told is pretty
> advanced) is investigating hybrid models of various dynamic
> programming algorithms to raise BLEU scores-- everyone's invested in
> the statistical (i.e. NOT interlingual) approach, and trying to figure
> out how to maximize performance. I think that there would be real
> advantages to using lojban as an interlingual medium, instead of
> essentially trying to imitate humans through machine learning, but
> when you have huge corpora of bilingual (or better) natural language
> data and virtually no bilingual corpora with lojban, it's just
> infeasible. There were a few attempts to use Esperanto for that
> purpose a while back (even before this paper), and no one seems to
> cite them, except as failures.
>
> Here's a really good up-to-date intro to MT:
> http://docs.google.com/Doc?docid=0AYZKIeNnTBe2ZGd4azRrZm1fNTI2ZnpnYmRrZ2g
>
> And a powerpoint version with lots of diagrams:
> http://people.csail.mit.edu/people/koehn/publications/tutorial2003.pdf
>
> mu'o mi'e .ku'us.
>
>
> On 2009-12-21, Jon "Top Hat" Jones <eyeonus@gmail.com> wrote:
>> I recently came across
>> this<http://www.lojban.org/files/why-lojban/mactrans.txt>paper, which
>> discusses various methods of machine translation methods. In it
>> it is mentioned that computers are not able to understand first-order
>> (i.e.
>> predicate) logic. Since the paper is nearly 2 decades old, I was
>> wondering
>> if anyone here knows what progress there has been in making it
>> understandable by computers.
>>
>>
>> --
>> mu'o mi'e .aionys.
>>
>> .i.a'o.e'e ko klama le bende pe denpa bu
>>
>

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.