Citation Extractor

1
Citation Extractor

• Nguyen Bach
• Sue Ann Hong
• Ben Lambert

2
Extraction Task
AuthorOf(Author, Paper) PublishedAt(Paper,
Conference) IsPaper, IsAuthor, IsConference

• Citation ltPaper, Authors, Conferencegt
• Pattern
• regular expression

3
Method Outline
Web pages (HTML, text)
Query Search (WIT)
Extract Patterns using known citations
Extract Citations using new patterns
Citations
Page-specific Patterns
4
Query "multiple-goal recognition from low-level
signals " " Xiaoyong Chai" " Qiang Yang" "AAAI
2005 "
Page http//www.informatik.uni-trier.de/ley/db/i
ndices/a-tree/y/YangQiang.html
5
Finding New Citations
6
The Challenge Patterns

• Beginning and the end
• Start token? End token? HTML tags?
• ?difficult to find length of token vs. general
• NER?
• These things should be talked about while viewing
  the previous slide
• Are regexs sufficient? (but not really relevant
  for self-supervised learning)
• Incorporating NER as a source of possible ENTITY
  marker?
• Use like AUTHOR, TITLE, CONF but with
  probabilities/confidence values

7
System Spits Out

• 6 seeds ? 60 citations
• 36 of these (partial citations)
• "Theory and Algorithms for Plan Merging " , "
  Ming Li"
• "The Expected Value of Hierarchical
  Problem-Solving " , " Fahiem Bacchus"
• "Handling feature interactions in
  process-planning "
• 14 of these (partial strings)
• "On D "
• "On t " , " John Tromp", " Elizabeth Sweedyk", "
  Umest Vazirani"
• "An L " , " Ronan Sleep"
• "To D
• No new conferences (end-token)

8
Bootstrapping, Short-Lived

• Highly restrictive regexs
• No recovery
• More seeds and variety the better
• Stupid Little Things
• Mis-capitalization
• Variations in titles (- vs. )
• Etc, etc, etc

9
Why is this one hard?
10
Extensions Improvements

• Less strict string matching
• Not case and punctuation sensitive
• Better boundary detection
• Start/end tokens, HTML wrapper detection?
• Better pattern construction
• e.g. n authors not 2
• NER
• help find the right "window
• A source of ENTITY marker
• Use like AUTHOR, TITLE, CONF but with
  probabilities/confidence values
• Evaluation with DBLP?

11
NER

• Baseline model (News corpus)
• ltENAMEX_TYPE"PERSON"gt M. Woszczyna, N.
  Aoki-Waibel, F. D. Buo, N. Coccaro, K. Horiguchi,
  T. Kemp, A. Lavie, A. McNair, T. Polzin, I.
  Rogina, C. P. Rose, T. Schultz, B. Suhm, M.
  Tomita, A. Waibel, 1994, JANUS 93 lt/ENAMEXgt
  Towards Spontaneous Speech Translation,
  Proceedings of the International Conference on
  Acoustics, Speech, and Signal Processing.
• ltENAMEX_TYPE"PERSON"gt S. Awodey. lt/ENAMEXgt
  Topological Representation of the Lambda
  Calculus. September ltENAMEX_TYPE"PERSON"gt 1998.
  Math. Struct. lt/ENAMEXgt in ltENAMEX_TYPE"LOCATION"
  gt Comp. Sci. (2000), vol. 10, pp. 81--96.
  lt/ENAMEXgt
• Adapted model (News citation corpus)
• ltENAMEX_TYPE"PERSON"gt M. Woszczyna, N.
  Aoki-Waibel, F. D. Buo, N. Coccaro, K. Horiguchi,
  T. Kemp, A. Lavie, A. McNair, T. Polzin, I.
  Rogina, C. P. Rose, T. Schultz, B. Suhm, M.
  Tomita, A. Waibel, 1994, JANUS 93 lt/ENAMEXgt
  Towards Spontaneous Speech Translation,
  Proceedings of the ltENAMEX_TYPE"ORGANIZATION"gt
  International Conference on Acoustics, Speech,
  lt/ENAMEXgt and Signal Processing.
• ltENAMEX_TYPE"PERSON"gt L. Birkedal. lt/ENAMEXgt A
  General Notion of Realizability. December 1999.
  Proceedings of ltENAMEX_TYPE"ORGANIZATION"gt LICS
  2000 lt/ENAMEXgt

12
NER

• HMM-based Model (Bikels 99)
• Baseline NER 94 F-score
• Trained 1.1 million words in News and
  Broadcastnews domain
• Apply Baseline Model to recognize
• Author, Conference , Location

13
NER Example with Baseline Model

• ltENAMEX_TYPE"PERSON"gt M. Woszczyna, N.
  Aoki-Waibel, F. D. Buo, N. Coccaro, K. Horiguchi,
  T. Kemp, A. Lavie, A. McNair, T. Polzin, I.
  Rogina, C. P. Rose, T. Schultz, B. Suhm, M.
  Tomita, A. Waibel, 1994, JANUS 93 lt/ENAMEXgt
  Towards Spontaneous Speech Translation,
  Proceedings of the International Conference on
  Acoustics, Speech, and Signal Processing.
• ltENAMEX_TYPE"PERSON"gt L. Birkedal. lt/ENAMEXgt A
  General Notion of Realizability. December 1999.
  Proceedings of LICS 2000
• ltENAMEX_TYPE"PERSON"gt S. Awodey. lt/ENAMEXgt
  Topological Representation of the Lambda
  Calculus. September ltENAMEX_TYPE"PERSON"gt 1998.
  Math. Struct. lt/ENAMEXgt in ltENAMEX_TYPE"LOCATION"
  gt Comp. Sci. (2000), vol. 10, pp. 81--96.
  lt/ENAMEXgt
• Good at detecting Author names boundaries, but
  sometimes too aggressive.

14
Adaptation NER

• Goals adapt baseline model to work better in
  citation domain.
• Issue No training data.
• A Solution Take 300 citations Run baseline
  model then recorrect them
• Train multiply 300 citations by 10, then train
  adaptation model with broadcast news corpus.

15
NER Example with Adaptation Model

• ltENAMEX_TYPE"PERSON"gt M. Woszczyna, N.
  Aoki-Waibel, F. D. Buo, N. Coccaro, K. Horiguchi,
  T. Kemp, A. Lavie, A. McNair, T. Polzin, I.
  Rogina, C. P. Rose, T. Schultz, B. Suhm, M.
  Tomita, A. Waibel, 1994, JANUS 93 lt/ENAMEXgt
  Towards Spontaneous Speech Translation,
  Proceedings of the ltENAMEX_TYPE"ORGANIZATION"gt
  International Conference on Acoustics, Speech,
  lt/ENAMEXgt and Signal Processing.
• ltENAMEX_TYPE"PERSON"gt L. Birkedal. lt/ENAMEXgt A
  General Notion of Realizability. December 1999.
  Proceedings of ltENAMEX_TYPE"ORGANIZATION"gt LICS
  2000 lt/ENAMEXgt
• ltENAMEX_TYPE"PERSON"gt D. Litman, D. Bhembe, C.
  P. Rose, K. Forbes-Riley, S. Silliman, K.
  VanLehn (2004). lt/ENAMEXgt Spoken Versus Typed
  Human and Computer Dialogue Tutoring, Proceedings
  of the Intelligent Tutoring Systems Conference.

16
How NER can help?

• Provide system generic Patterns.
• AUTHOR M. Woszczyna, N. Aoki-Waibel, F. D. Buo,
  N. Coccaro, K. Horiguchi, T. Kemp, A. Lavie, A.
  McNair, T. Polzin, I. Rogina, C. P. Rose, T.
  Schultz, B. Suhm, M. Tomita, A. Waibel, 1994,
  JANUS 93
• CONFERENCE International Conference on
  Acoustics, Speech
• Then use specific rules to refine

17
Lessons LearnedAnother Boring Text Slide

• Semi-structured text is surprisingly difficult to
  read
• Off-line training for wrappers and/or NER may
  help
• Need very high-confidence rules to ensure
  precision
• A continuously-running system needs robustness
  (internet/Google-failure, unexpected errors, )