Does anyone use "discounts" to adjust the zero
probability of prolegomena(Sp?) Marina Santini's
course goes into great detail on several discount
theories. They all involve fudging the basic
statistical number, and she says that is because
it "performs better with the unseen cases", by
which she means unseen in the training class, but
likely to appear later with more material. (01)
-Rich (02)
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2 (03)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On
Behalf Of Rich Cooper
Sent: Saturday, September 27, 2014 5:51 PM
To: '[ontolog-forum] '
Subject: [ontolog-forum] Marina Santini's course
in Linguistics (04)
I found a course (by Marina Santini in Sweden's
Uppsala University). Her home page shows the
first few lectures, along with many other items of
linguistic interest. But I found 7 lecture notes
which she posted to explain the math very clearly. (05)
Here are the URLs:
Marina's Lecture Slides on NLP and Corpus Analysis
--------------------------------------------------
-------
http://www.slideshare.net/marinasantini1/lecture-1
-introduction-tothecourse
http://www.slideshare.net/marinasantini1/lecture-2
-fromsemanticstoapplications
http://www.slideshare.net/marinasantini1/lecture-3
-probabilitytheory
http://www.slideshare.net/marinasantini1/lecture2-
statistical-inference
http://www.slideshare.net/fullscreen/marinasantini
1/lecture-5-bayesian-classification/1
http://www.slideshare.net/fullscreen/gendrong/1ed-
jurafsky-martin-chap6ngram/1
http://www.slideshare.net/marinasantini1/lecture-6
-39243943
http://www.slideshare.net/marinasantini1/lecture-7
-hidden-markov-models-hmms (06)
I think she is going to offer more lectures before
the course is over. But for those seeking a
simple equation description of linguistic filters
and predictive functions, she provides useful
guidance, including the need for "discounts" to
offset zero count N-grams and such adjustments to
the usual statistical models. (07)
-Rich (08)
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On
Behalf Of John Bottoms
Sent: Saturday, September 20, 2014 8:10 AM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Generalization,
specialization, and interoperability (09)
John, Ed, (010)
I view it as a typical process in specialization.
First the containers
are set up and then the types of contents are
specified, finally the
contents are added. EFT is more open world now,
and allows for SGML
documents (record 914 if I recall). (011)
-John Bottoms (012)
On 9/19/2014 10:25 PM, Rich Cooper wrote:
> Dear John, Ed et al,
>
> John wrote:
> the EFT ontology is the very
> underspecified common theory on which
> interoperability is possible.
>
> Why just 'very underspecified'? Wouldn't it be
> 'minimally underspecified'? There is no other
> knowledge built into EFT other than how to
> transfer money electronically, so bank account
> numbers, routing numbers, and so forth are
> specified in EFT. So adding anything at all to
> EFT would be nonminimal.
>
> -Rich
>
> Sincerely,
> Rich Cooper
> EnglishLogicKernel.com
> Rich AT EnglishLogicKernel DOT com
> 9 4 9 \ 5 2 5 - 5 7 1 2
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx]
On
> Behalf Of John F Sowa
> Sent: Friday, September 19, 2014 6:48 PM
> To: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] Generalization,
> specialization, and interoperability
>
> David and Ed,
>
> The criteria I mentioned only work for two
systems
> A and B whose
> ontologies have *both* been defined as
> specializations of a common
> generalization C.
>
> DW
>> On the face of it, these [Acme and Enterprise]
> are both compatible
>
> The criteria I stated do not include a condition
> called "face of it".
> Following is a copy from my previous note:
>
> JFS
>> a) If two theories A and B are inconsistent in
> their details,
>> they can interoperate on shared data that
is
> specified by
>> a common generalization C.
>>
>> b) To use data specified in C, neither A nor B
> may assume any
>> properties of that data not specified in C.
> But they can use
>> the details in conditionals that begin "If
x
> has property P ..."
>
> If the designers of Acme and Enterprise had both
> adopted some common
> standard C, then they could interoperate under
> these conditions.
> But if they hadn't, then there are two options:
>
> 1. Give up.
>
> 2. Analyze the spec's for Acme and
Enterprise
> to determine whether
> there exists a common generalization for
> some subset of the data.
>
> EJB
>> For the ontology containing only these
> predicates, we have no easy
>> way to know what satisfying these axioms might
> be.
>
> I agree.
>
> For practical examples, note that many standards
> have evolved over
> the years to support electronic funds transfer
> (EFT). But those
> standards specify almost nothing about the
> accounts from which
> and to which the funds are being transferred.
>
> Over the years, banks have merged with or
> swallowed other banks many,
> many times. They all have accounts with similar
> sounding names:
> checking, savings, etc. But the banks never,
ever
> merge accounts --
> for reasons along the lines that David and Ed
have
> summarized.
> Instead, they have two options:
>
> 1. Give up -- i.e., continue to run the old
> software from both banks
> indefinitely.
>
> 2. Shut down some or all the old accounts,
open
> similar sounding
> accounts that are processed by different
> software, and transfer
> the funds from the old accounts to the new
> accounts.
>
> To relate this example to my criteria above, the
> EFT ontology is
> the very underspecified common theory on which
> interoperability
> is possible.
>
> John
> (013)
__________________________________________________
_______________
Message Archives:
http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr:
http://ontolog.cim3.net/mailman/listinfo/ontolog-f
orum/
Unsubscribe:
mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join:
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePa
ge#nid1J (014)
__________________________________________________
_______________
Message Archives:
http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr:
http://ontolog.cim3.net/mailman/listinfo/ontolog-f
orum/
Unsubscribe:
mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join:
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePa
ge#nid1J (015)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (016)
|