fine grained evaluation for entity linking
play

Fine-Grained Evaluation for Entity Linking Henry Rosales-M endez, - PowerPoint PPT Presentation

Fine-Grained Evaluation for Entity Linking Henry Rosales-M endez, Aidan Hogan and Barbara Poblete University of Chile { hrosales,ahogan,bpoblete } @dcc.uchile.cl Nov 5, 2019 EMNLP-IJCNLP 2019, Hong Kong Example Example - Entity


  1. Fine-Grained Evaluation for Entity Linking Henry Rosales-M´ endez, Aidan Hogan and Barbara Poblete University of Chile { hrosales,ahogan,bpoblete } @dcc.uchile.cl Nov 5, 2019 † EMNLP-IJCNLP 2019, Hong Kong

  2. Example

  3. Example - Entity Recognition

  4. Example - Entity Disambiguation

  5. Knowledge Bases

  6. Name Variations in Entity Linking Michael Joseph Jackson Michael J. Jackson King of Pop

  7. Name Variations in Entity Linking Michael Jackson

  8. Text collections with noise

  9. Text collections with noise

  10. Text collections with noise

  11. Multilingual Entity Linking - English

  12. Multilingual Entity Linking - Italian

  13. Multilingual Entity Linking - Spanish

  14. • What should Entity Linking link?

  15. Example annotations produced by four EL systems

  16. Example annotations produced by four EL systems

  17. Example annotations produced by four EL systems Aida

  18. Example annotations produced by four EL systems Babelfy Aida

  19. Example annotations produced by four EL systems Babelfy Aida DBpedia Spotlight

  20. Example annotations produced by four EL systems Babelfy Aida T agME DBpedia Spotlight

  21. Questionnaire Gong-Qing Wu, Ying He, and Xuegang Hu. 2018. Entity linking: An issue to extract corresponding entity with knowledge base. IEEE Access, 6:62206231.

  22. Questionnaire In an interview with Martin Bashir for the 2003 documentary Living with Michael Jackson, the King of Pop recalled that Joe often sat with a white belt at hand as he and his four siblings rehearsed. Russian daily Kommersant reports that Moscow will supply the Greeks with gas at rock bottom prices as Tsipras prepares to meet the Russian President.

  23. Questionnaire

  24. Questionnaire 1 We sent the questionnaire to 321 researchers, of which 232 requests were delivered successfully. 2 We received a total of 36 responses.

  25. Questionnaire

  26. Questionnaire

  27. Goal 1 Allowing a fine-grained evaluation for Entity Linking

  28. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  29. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  30. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  31. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  32. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  33. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  34. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  35. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  36. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  37. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  38. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  39. Categories B ASE F ORM P ART OF S PEECH O VERLAP R EFERENCE Proper Noun Noun Phrase None Direct Full Name Singular Maximal Anaphoric Short Name Plural Intermediate Metaphoric Extended Name Adjective Minimal Metonymic Alias Verb Related Numeric/Temporal Adverb Descriptive Common Form Pro-Form

  40. Re-annotation and categorization 1 KORE50: 1 doc, 50 sentences 2 VoxEL: 15 doc, 94 sentences 3 ACE04: first 20 doc, 214 sentences

  41. Re-annotation and categorization

  42. Re-annotation and categorization

  43. Validating by tag

  44. Validating by tag

  45. Validating by tag

  46. How to perform an evaluation for categorized datasets?

  47. Traditional F1 P = | TP | | S | R = | TP | | G | F 1 = 2 · P · R P + R

  48. Fuzzy set

  49. Modifications to F1 P = | TP | | S | � a ∈ S µ G ∗ ( a ) R ∗ = � a ∈ G µ G ∗ ( a ) F 1 = 2 · P · R ∗ P + R ∗

  50. Prop1 : the values for R ∗ and F ∗ • 1 both range between 0 and 1, inclusive. • Prop2 : when µ G ∗ : G → { 1 } (i.e., when memberships are binary), R ∗ and F ∗ 1 correspond to R and F 1 . • Prop3 : missing annotations with higher membership degree are penalized more in R ∗ and F ∗ 1 than those with lower degree.

  51. Conclusions • We stress the lack of consensus about what should Entity Linking link. • We propose a set of categories for Entity Linking. • We re-annotate three datasets: VoxEL, KORE50 and ACE04 − first 20. • We extend F 1 measure.

  52. Fine-Grained Evaluation for Entity Linking Henry Rosales-M´ endez, Aidan Hogan and Barbara Poblete University of Chile { hrosales,ahogan,bpoblete } @dcc.uchile.cl Nov 5, 2019 † EMNLP-IJCNLP 2019, Hong Kong

Recommend


More recommend