[Corpora-List] Wavelet for NLP

Pascale Fung pascale at cs.ust.hk
Mon Jun 12 11:08:00 CEST 2006


Dear Stefan and Bill,

Since Bill asked about wavelet in NLP applications, I don't think he has
coding in mind particularly. Correct me if I am wrong. I would also be
interested in knowing what Bill and others on the list might have in mind
for using wavelet in NLP.

Regards,
Pascale

----------------------------------------------------------------------------
Prof. Pascale Fung
Department of Electronic & Computer Engineering
University of Science & Technology
Clear Water Bay, Kowloon
Hong Kong

http://www.ee.ust.hk/~pascale
pascale at ee.ust.hk
tel:+852 2358 8537
fax:+852 2358 1485
sec:+852 2358 7087
----------------------------------------------------------------------------


> Thanks a lot for the detailed clarification. I've always been

> thinking of wavelet transforms as a "variant" of Fourier

> transformation, which is also (at least supposed to be) invertible in

> the continuous case.

>

> My impression from the original query was that the author is more

> interested in using for coding or data manipulation rather than just

> analysis, but this may be purely due to my Fourier-based

> perspective. :-)

>

> Best,

> Stefan

>

> On 10 Jun 2006, at 19:18, Pascale Fung wrote:

>

>> "Time frequency transformation" is basically wavelet transform.

>>

>> I think you are talking about discrete wavelet transform, which is

>> bijective, and used for source coding purposes. I used continuous

>> wavelet

>> transform, which is injective, and used for recognition (or analysis)

>> purposes.

>>

>> Discrete wavelet transform is used for coding purposes where you'd be

>> concerned with recovering the original signal. Whereas in the

>> application

>> of bilingual word translation, I was interested in recognizing the

>> patterns. I would say most NLP tasks are recognition rather than

>> coding

>> tasks.

>>

>> Nevertheless, in this particular recognition application (of bilingual

>> word pair extraction) you can still recover the orginal "signal"

>> from the

>> output of the transformation because the output can only correspond

>> to one

>> and only one input.

>>

>> regards,

>> Pascale

>>

>

>








More information about the Corpora-archive mailing list