1 / 84

Features & Unification

Features & Unification. Ling 571 Deep Processing Techniques for NLP January 31, 2011. Roadmap. Features: Motivation Constraint & compactness Features Definitions & representations Unification Application of features in the grammar Agreement, subcategorization

xanti
Download Presentation

Features & Unification

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Features & Unification Ling 571 Deep Processing Techniques for NLP January 31, 2011

  2. Roadmap • Features: Motivation • Constraint & compactness • Features • Definitions & representations • Unification • Application of features in the grammar • Agreement, subcategorization • Parsing with features & unification • Augmenting the Earley parser, unification parsing • Extensions: Types, inheritance, etc • Conclusion

  3. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs.

  4. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs. • But… • *They runs • *He run • *He disappeared the flight

  5. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs. • But… • *They runs • *He run • *He disappeared the flight • NP -> Det Nom • This flight

  6. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs. • But… • *They runs • *He run • *He disappeared the flight • NP -> Det Nom • This flight • These flights

  7. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs. • But… • *They runs • *He run • *He disappeared the flight • NP -> Det Nom • This flight • These flights • *This flights

  8. Constraints & Compactness • Constraints in grammar • S -> NP VP • They run. • He runs. • But… • *They runs • *He run • *He disappeared the flight • NP -> Det Nom • This flight • These flights • *This flights • Violate agreement (number), subcategorization

  9. Enforcing Constraints • Enforcing constraints

  10. Enforcing Constraints • Enforcing constraints • Add categories, rules

  11. Enforcing Constraints • Enforcing constraints • Add categories, rules • Agreement: • S-> NPsg3p VPsg3p, • S-> NPpl3p VPpl3p,

  12. Enforcing Constraints • Enforcing constraints • Add categories, rules • Agreement: • S-> NPsg3p VPsg3p, • S-> NPpl3p VPpl3p, • Subcategorization: • VP-> Vtrans NP, • VP -> Vintrans, • VP->Vditrans NP NP

  13. Enforcing Constraints • Enforcing constraints • Add categories, rules • Agreement: • S-> NPsg3p VPsg3p, • S-> NPpl3p VPpl3p, • Subcategorization: • VP-> Vtrans NP, • VP -> Vintrans, • VP->Vditrans NP NP • Explosive!, loses key generalizations

  14. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is

  15. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is • number: sg, pl • I am; we are

  16. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is • number: sg, pl • I am; we are • case: nom, acc • I, he; me, him • gender: masc, fem, neut • animacy: +/1 • etc

  17. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is • number: sg, pl • I am; we are • case: nom, acc • I, he; me, him

  18. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is • number: sg, pl • I am; we are • case: nom, acc • I, he; me, him • gender: masc, fem, neut

  19. Features • person: 1st, 2nd, 3rd • I, we; you; he, she, they • am, are, is • number: sg, pl • I am; we are • case: nom, acc • I, he; me, him • gender: masc, fem, neut • animacy: +/1 • etc

  20. Why features? • Need compact, general constraints • S -> NP VP

  21. Why features? • Need compact, general constraints • S -> NP VP • Only if NP and VP agree

  22. Why features? • Need compact, general constraints • S -> NP VP • Only if NP and VP agree • How can we describe agreement, subcat?

  23. Why features? • Need compact, general constraints • S -> NP VP • Only if NP and VP agree • How can we describe agreement, subcat? • Decompose into elementary features that must be consistent • E.g. Agreement

  24. Why features? • Need compact, general constraints • S -> NP VP • Only if NP and VP agree • How can we describe agreement, subcat? • Decompose into elementary features that must be consistent • E.g. Agreement • Number, person, gender, etc

  25. Why features? • Need compact, general constraints • S -> NP VP • Only if NP and VP agree • How can we describe agreement, subcat? • Decompose into elementary features that must be consistent • E.g. Agreement • Number, person, gender, etc • Augment CF rules with feature constraints • Develop mechanism to enforce consistency • Elegant, compact, rich representation

  26. Feature Representations • Fundamentally, Attribute-Value pairs • Values may be symbols or feature structures • Feature path: list of features in structure to value • “Reentrant feature structures”: share some struct • Represented as • Attribute-value matrix (AVM), or • Directed acyclic graph (DAG)

  27. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set

  28. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set Attribute-value matrix (AVM)

  29. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set Attribute-value matrix (AVM) NUMBER PL

  30. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set Attribute-value matrix (AVM) NUMBER PL PERSON 3

  31. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set Attribute-value matrix (AVM) NUMBER PL PERSON 3 NUMBER PL PERSON 3

  32. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set Attribute-value matrix (AVM) NUMBER PL PERSON 3 NUMBER PL PERSON 3 CAT NP NUMBER PL PERSON 3

  33. Feature Representations • Fundamentally, Attribute-Value pairs • Features: atomic symbols from a finite set • Values may be • Atomic symbols from a finite set • Values may also be feature structures themselves Attribute-value matrix (AVM) CAT NP AGREEMENT NUMBER PL PERSON 3

  34. Feature Representations • Feature path: • Sequence of features through a feature structure leading to a particular value CAT NP AGREEMENT NUMBER PL PERSON 3

  35. Feature Representations • Feature path: • Sequence of features through a feature structure leading to a particular value CAT NP AGREEMENT NUMBER PL PERSON 3 <AGREEMENT NUMBER> -> PL

  36. Feature Representations • Feature path: • Sequence of features through a feature structure leading to a particular value CAT NP AGREEMENT NUMBER PL PERSON 3 <AGREEMENT NUMBER> -> PL <AGREEMENT PERSON> -> 3

  37. Feature Representations • Reentrant feature structures • Features share some feature structure as value • Not merely equal values • Shared substructure • Feature paths lead to same node CAT S HEAD AGREEM’T NUMBER PL PERSON 3 1 1 SUBJECT AGREEMENT

  38. AVM CAT NP AGREEMENT NUMBER PL NUMBER PL PERSON 3 PERSON 3 NUMBER PL PERSON 3 CAT S HEAD AGREEM’T NUMBER PL PERSON 3 1 CAT NP NUMBER PL PERSON 3 1 SUBJECT AGREEMENT

  39. Head-Subject Agreement CAT S HEAD AGREEM’T NUMBER PL PERSON 3 1 1 SUBJECT AGREEMENT

  40. Feature representations • Feature structures can also be represented as DAGs • Directed, acyclic graphs • Edges are features • Nodes values

  41. Reentrant DAG

  42. Unification • Two key roles:

  43. Unification • Two key roles: • Merge compatible feature structures

  44. Unification • Two key roles: • Merge compatible feature structures • Reject incompatible feature structures

  45. Unification • Two key roles: • Merge compatible feature structures • Reject incompatible feature structures • Two structures can unify if

  46. Unification • Two key roles: • Merge compatible feature structures • Reject incompatible feature structures • Two structures can unify if • Feature structures are identical • Result in same structure

  47. Unification • Two key roles: • Merge compatible feature structures • Reject incompatible feature structures • Two structures can unify if • Feature structures are identical • Result in same structure • Feature structures match where both have values, differ in missing or underspecified • Resulting structure incorporates constraints of both

  48. Subsumption • Relation between feature structures • Less specific f.s. subsumes more specific f.s. • F.s. F subsumes f.s. G iff • For every feature x in F, F(x) subsumes G(x) • For all paths p and q in F s.t. F(p)=F(q), G(p)=G(q)

  49. Subsumption • Relation between feature structures • Less specific f.s. subsumes more specific f.s. • F.s. F subsumes f.s. G iff • For every feature x in F, F(x) subsumes G(x) • For all paths p and q in F s.t. F(p)=F(q), G(p)=G(q) • Examples: • A: [Number SG], B: [Person 3] • C:[Number SG] • [Person 3]

  50. Subsumption • Relation between feature structures • Less specific f.s. subsumes more specific f.s. • F.s. F subsumes f.s. G iff • For every feature x in F, F(x) subsumes G(x) • For all paths p and q in F s.t. F(p)=F(q), G(p)=G(q) • Examples: • A: [Number SG], B: [Person 3] • C:[Number SG] • [Person 3] • A subsumes C

More Related