On 11/15/2014 11:37 AM, Eric Prud'hommeaux wrote:
> I interpret "fragmentation" as failure to re-use existing work
> when possible. (01)
I meant something more than just reusing software. I meant that
it's essential to integrate multiple *paradigms*. In my paper,
I mentioned formal logic, neural networks, NL parsers, analogy,
and any other *paradigm* anybody has ever thought of. (02)
As an example, I discussed Watson's approach to Jeopardy!. There
were over 30 researchers who contributed to the papers in the
IBM journal issue. And the methods they used ranged over a very
wide range of paradigms: ontology, NL parsing, machine learning,
large knowledge sources, Wikipedia, DBpedia, statistics of various
kinds, pattern matching, formal and informal reasoning, etc., etc. (03)
Watson certainly took advantage of the sources available on the
SW, but the methods they developed go very far beyond the usual
way that people use RDF + OWL + IRIs. (04)
Tim B-L said that the SW is *not* AI. And I agree. My article
emphasizes that there is a huge amount of AI that isn't being
used. I don't recommend that the average programmer should
attempt to use those methods. But it's essential for the
researchers to develop tools that can enable ordinary people
to take advantage of *all* those paradigms. Unfortunately,
most researchers just stick with one or two familiar paradigms. (05)
As an example of a challenge for AI (including ontology),
Hector Levesque cited "Winograd Schemas": (06)
http://www.cs.toronto.edu/~hector/Papers/winograd.pdf (07)
And see below for some examples of Winograd-style sentences. (08)
Any system that can answer Jeopardy! questions must be deal
with such issues. In fact, any system that can analyze any NL
text on any subject must deal with them. (09)
John
__________________________________________________________________ (010)
Source: https://www.cs.nyu.edu/davise/papers/WS.html (011)
A Collection of Winograd Schemas (012)
A Winograd schema is a pair of sentences that differ in only one or two
words and that contain an ambiguity that is resolved in opposite ways in
the two sentences and requires the use of world knowledge and reasoning
for its resolution. The schema takes its name from a well-known example
by Terry Winograd (1972): (013)
"The city councilmen refused the demonstrators a permit because they
[feared/advocated] violence." (014)
If the word is 'feared', then 'they' presumably refers to the city
council; if it is ``advocated'' then ``they'' presumably refers to the
demonstrators. (015)
[Note by JFS: Many linguists and AI researchers had cited
similar examples. They are common in ordinary language.] (016)
In his paper, "The Winograd Schema Challenge" Hector Levesque (2011)
proposes to assemble a set of such Winograd schemas that are (017)
* Easily disambiguated by the human reader (ideally, so easily that
the reader does not even notice that there is an ambiguity); (018)
* Not solvable by simple techniques such as selectional restrictions; (019)
* Google-proof; that is, there is no obvious statistical test over
text corpora that will reliably disambiguate these correctly.
... (020)
First five of 100+ Winograd Schemas: (021)
1. The city councilmen refused the demonstrators a permit because they
[feared/advocated] violence. Who [feared/advocated] violence?
Answers: The city councilmen/the demonstrators. (022)
2. The trophy doesn't fit into the brown suitcase because it's too
[small/large]. What is too [small/large]?
Answers: The suitcase/the trophy. (023)
3. Joan made sure to thank Susan for all the help she had
[given/received]. Who had [given/received] help?
Answers: Susan/Joan. (024)
4. Paul tried to call George on the phone, but he wasn't
[successful/available]. Who was not [successful/available]?
Answers: Paul/George (025)
5. The lawyer asked the witness a question, but he was reluctant to
[answer/repeat] it. Who was reluctant to [answer/repeat] it?
Answers: The witness/the lawyer. (026)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2015/
Community Wiki: http://ontolog-02.cim3.net/wiki/OntologySummit2015
Community Portal: http://ontolog-02.cim3.net/wiki/ (027)
|