Extracting Paper Titles, Authors and Conferences from Lists on the Web

1
Extracting Paper Titles, Authors and Conferences
from Lists on the Web

• Nguyen Bach
• Sue Ann Hong
• Ben Lambert

2
We will attempt to extract these predicates and
relations

• isAuthor(X)
• isPaperTitle(X)
• isConferenceName(X)
• publishedAt( ltpapergt, ltconferencegt )
• authorOf (ltauthorgt , ltpapergt )

3
By looking at these Web pages

• An authors publication lists (on their home
  page)
• Conference accepted papers list (on the
  conference Web page)

4
AAAI-05 Paper List
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Then we look for patterns

• ltauthorgt and ltauthorgt, lttitlegt.
• ltauthorgt , ltauthorgt, ltauthorgt and ltauthorgt,
  lttitlegt.
• ltauthorgt , ltauthorgt, ltauthorgt and ltauthorgt,
  lttitlegt.

11
But

• Maybe the patterns are wrong, so look for some
  more evidence
• Once we have enough evidence we can add it to our
  knowledge base.

12
And around 100 other citations in roughly the
same format
13
Conclusion

• Redundancy of authors 1
• Seems to be very precise since titles usually do
  not have many spelling variations.
• Help to find alternate name spellings
• Could there be a Q. Yang and a Qiang Yang
  with nearly identical publications?
• Works for other fields also (e.g. History)