Chinese Firewall Update Leif Guillermo, Veronika Strnadova This material is based upon work supported by the National Science Foundation under Grant No.

Chinese Firewall Update

Leif Guillermo, Veronika Strnadova

This material is based upon work supported by the National Science Foundation under Grant No. IIS/REU/0755462.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not

necessarily reflect the views of the National Science Foundation.

Chinese Firewall Update Overview Long-term goal The Present Situation Development up to present Potential Alternatives 'What To Do'

http://chinadigitaltimes.net/wp-content/uploads/2008/07/great_firewall.jpg

Overview Want:

An extensive list of possibly blacklisted keywords

A better understanding of the GFC and

internet censorship. Method(s):

Chinese Character Phonological

Similarities Chinese Character Recognition

Long-term goal The goal is to create an extensive list

of possibly blacklisted keywords Manual Labor vs. Automation with

Adaptation The Chinese Firewall, Censorship,

and you

The Present Situation Reaching Out

We've spoken with one professor and are

scheduled to speak with others Currently Working on Phonology

Mappings:

Character->Sound Sound->Character

Permutations

Character recognition still a go

Image Feature Extraction SVD

http://www.itanlp.com/images/NLPdiag.gif

Mapping Example

http://www.ofoghlu.net/log/Lanet-vi-Internet-Map.jpg

Development up to present We've read some good papers The Time Course of Graphic, Phonological, and Semantic Activation in

Chinese Character Identification. Charles A. Perfetti and Li Hai Tan. Journal of Experimental Psychology: Learning Memory, and Cognition 1998, Vol. 24, No. 1, 101-118

Uncalibrated Stereo Correspondence by Singular Value Decomposition. Maurizio Pilu. Digital Media Department. HP Laboratories Bristol. HPL-97-96. August, 1997

Originally wanted CCR, ended up with

CPAs Original CCR method should still be useful

Potential Alternatives Phonetic Alignment and Similarity

Could be more rigorous than the current phonological method but need to do more research.

Gabor Filters on different parts of Chinese

Characters.

Could potentially be as accurate if not more

accurate than Feature Extraction. Possibly not what we want. Complicated, and very math intensive.

'What to do' Currently:

Graphical<->Phonological Mappings are

almost done. Next Step:

Create an algorithm to generate possible word

permutations. Tasks:

Veronika is working on the permutation algorithm. I've been dealing with the mappings. We've both been reading/gathering information.

Anyone want to recommend one or several strategies?