delexicalized parsing
play

Delexicalized Parsing Daniel Zeman, Rudolf Rosa April 3, 2020 - PowerPoint PPT Presentation

Feb 24, 2024 •22 likes •622 views

Delexicalized Parsing Daniel Zeman, Rudolf Rosa April 3, 2020 NPFL120 Multilingual Natural Language Processing Charles University Faculty of Mathematics and Physics Institute of Formal and Applied Linguistics unless otherwise stated

  1. Delexicalized Parsing Daniel Zeman, Rudolf Rosa April 3, 2020 NPFL120 Multilingual Natural Language Processing Charles University Faculty of Mathematics and Physics Institute of Formal and Applied Linguistics unless otherwise stated

  2. Delexicalized Parsing MD VB CC VB IN DT NN Delexicalized Parsing 1/22 • What if we feed the parser with tags instead of words? • Ændringer i listen i bilaget ofgentliggøres og meddeles på samme måde. • NNS IN NN IN NN VB CC VB IN DT NN • NNS IN NN • Förändringar i förteckningen skall ofgentliggöras och meddelas på samma sätt.

  3. Delexicalized Parsing ((MD (VB CC VB)) (IN (DT NN)))) Delexicalized Parsing 2/22 • What if we feed the parser with tags instead of words? • Ændringer i listen i bilaget ofgentliggøres og meddeles på samme måde. • ((NNS (IN NN (IN NN))) ((VB CC VB) (IN (DT NN)))) • ((NNS (IN NN)) • Förändringar i förteckningen skall ofgentliggöras och meddelas på samma sätt.

  4. • (JRC-Acquis parallel corpus) • Hajič tagger for Swedish (PAROLE tagset) • CoNLL 2006 treebanks (dependencies) • Danish Dependency Treebank • Swedish Talbanken05 • Two constituency parsers: • “Charniak” • “Brown” (Charniak N-best parser + Johnson reranker) • Other resources Danish – Swedish Setup Related Languages India Delexicalized Parsing 3/22 • Daniel Zeman, Philip Resnik (2008). Cross-Language Parser Adaptation between • In IJCNLP 2008 Workshop on NLP for Less Privileged Languages, pp. 35–42, Hyderabad,

  5. Danish – Swedish Setup Related Languages India Delexicalized Parsing 3/22 • Daniel Zeman, Philip Resnik (2008). Cross-Language Parser Adaptation between • In IJCNLP 2008 Workshop on NLP for Less Privileged Languages, pp. 35–42, Hyderabad, • CoNLL 2006 treebanks (dependencies) • Danish Dependency Treebank • Swedish Talbanken05 • Two constituency parsers: • “Charniak” • “Brown” (Charniak N-best parser + Johnson reranker) • Other resources • (JRC-Acquis parallel corpus) • Hajič tagger for Swedish (PAROLE tagset)

  6. • (JRC-Acquis parallel corpus) Danish – Swedish Setup Related Languages India Delexicalized Parsing 3/22 • Daniel Zeman, Philip Resnik (2008). Cross-Language Parser Adaptation between • In IJCNLP 2008 Workshop on NLP for Less Privileged Languages, pp. 35–42, Hyderabad, • CoNLL 2006 treebanks (dependencies) • Danish Dependency Treebank • Swedish Talbanken05 • Two constituency parsers: • “Charniak” • “Brown” (Charniak N-best parser + Johnson reranker) • Other resources • Hajič tagger for Swedish (PAROLE tagset)

  7. • NUM governs NOUN • NOUN governs NUM • GEN governs NOM • NOM governs GEN • COORD: last member on • COORD: member on Treebank Normalization Delexicalized Parsing and conjs on next member previous member, commas year’s income års inkomster Swedish and ADJ Danish on fjrst member conjunction, everything else Russia’s way Ruslands vej ADJ governs NOUN 4/22 • DET governs ADJ • NOUN governs both DET

  8. • GEN governs NOM • NOM governs GEN • COORD: last member on • COORD: member on Treebank Normalization and ADJ Delexicalized Parsing and conjs on next member previous member, commas year’s income års inkomster Swedish Danish on fjrst member conjunction, everything else Russia’s way Ruslands vej ADJ governs NOUN 4/22 • DET governs ADJ • NOUN governs both DET • NUM governs NOUN • NOUN governs NUM

  9. • COORD: last member on • COORD: member on Treebank Normalization Danish Delexicalized Parsing and conjs on next member previous member, commas year’s income års inkomster and ADJ Swedish on fjrst member conjunction, everything else Russia’s way Ruslands vej ADJ governs NOUN 4/22 • DET governs ADJ • NOUN governs both DET • NUM governs NOUN • NOUN governs NUM • GEN governs NOM • NOM governs GEN

  10. Treebank Normalization on fjrst member Delexicalized Parsing and conjs on next member previous member, commas year’s income års inkomster and ADJ Danish Swedish conjunction, everything else Russia’s way Ruslands vej ADJ governs NOUN 4/22 • DET governs ADJ • NOUN governs both DET • NUM governs NOUN • NOUN governs NUM • GEN governs NOM • NOM governs GEN • COORD: last member on • COORD: member on

  11. • Convert dependencies to constituents • Flattest possible structure • DA/SV tagset converted to Penn Treebank tags • Nonterminal labels: • derived from POS tags • then translated to the Penn set of nonterminals • Make the parser feel it works with the Penn Treebank • (Although it could have been confjgured to use other sets of labels.) Treebank Preparation Delexicalized Parsing 5/22 • Transform Danish to Swedish tree style • A few heuristics • Only for evaluation! Not needed in real world.

  12. • DA/SV tagset converted to Penn Treebank tags • Nonterminal labels: • derived from POS tags • then translated to the Penn set of nonterminals • Make the parser feel it works with the Penn Treebank • (Although it could have been confjgured to use other sets of labels.) Treebank Preparation Delexicalized Parsing 5/22 • Transform Danish to Swedish tree style • A few heuristics • Only for evaluation! Not needed in real world. • Convert dependencies to constituents • Flattest possible structure

  13. • Nonterminal labels: • derived from POS tags • then translated to the Penn set of nonterminals • Make the parser feel it works with the Penn Treebank • (Although it could have been confjgured to use other sets of labels.) Treebank Preparation Delexicalized Parsing 5/22 • Transform Danish to Swedish tree style • A few heuristics • Only for evaluation! Not needed in real world. • Convert dependencies to constituents • Flattest possible structure • DA/SV tagset converted to Penn Treebank tags

  14. Delexicalized Parsing Treebank Preparation 5/22 • Transform Danish to Swedish tree style • A few heuristics • Only for evaluation! Not needed in real world. • Convert dependencies to constituents • Flattest possible structure • DA/SV tagset converted to Penn Treebank tags • Nonterminal labels: • derived from POS tags • then translated to the Penn set of nonterminals • Make the parser feel it works with the Penn Treebank • (Although it could have been confjgured to use other sets of labels.)

  15. • da-sv lexicalized: Charniak = 43.28, Brown = 41.84 • (no morphology tweaking) • da-da delexicalized: Charniak = 79.62, Brown = 80.20 (!) • (hybrid sv-da Hajič-like tagset = “words”, Penn POS = “tags”) • sv-sv delexicalized: Charniak = 76.07, Brown = 77.01 • da-sv delexicalized: Charniak = 65.50, Brown = 66.40 Unlabeled F Scores Delexicalized Parsing 6/22 • da-da lexicalized: Charniak = 78.16, Brown = 78.24 • (CoNLL train 94K words, test 5852 words) • sv-sv lexicalized: Charniak = 77.81, Brown = 78.74 • (CoNLL train 191K words, test 5656 words)

  16. • da-da delexicalized: Charniak = 79.62, Brown = 80.20 (!) • (hybrid sv-da Hajič-like tagset = “words”, Penn POS = “tags”) • sv-sv delexicalized: Charniak = 76.07, Brown = 77.01 • da-sv delexicalized: Charniak = 65.50, Brown = 66.40 Unlabeled F Scores Delexicalized Parsing 6/22 • da-da lexicalized: Charniak = 78.16, Brown = 78.24 • (CoNLL train 94K words, test 5852 words) • sv-sv lexicalized: Charniak = 77.81, Brown = 78.74 • (CoNLL train 191K words, test 5656 words) • da-sv lexicalized: Charniak = 43.28, Brown = 41.84 • (no morphology tweaking)

  17. • sv-sv delexicalized: Charniak = 76.07, Brown = 77.01 • da-sv delexicalized: Charniak = 65.50, Brown = 66.40 Unlabeled F Scores Delexicalized Parsing 6/22 • da-da lexicalized: Charniak = 78.16, Brown = 78.24 • (CoNLL train 94K words, test 5852 words) • sv-sv lexicalized: Charniak = 77.81, Brown = 78.74 • (CoNLL train 191K words, test 5656 words) • da-sv lexicalized: Charniak = 43.28, Brown = 41.84 • (no morphology tweaking) • da-da delexicalized: Charniak = 79.62, Brown = 80.20 (!) • (hybrid sv-da Hajič-like tagset = “words”, Penn POS = “tags”)

  18. • da-sv delexicalized: Charniak = 65.50, Brown = 66.40 Unlabeled F Scores Delexicalized Parsing 6/22 • da-da lexicalized: Charniak = 78.16, Brown = 78.24 • (CoNLL train 94K words, test 5852 words) • sv-sv lexicalized: Charniak = 77.81, Brown = 78.74 • (CoNLL train 191K words, test 5656 words) • da-sv lexicalized: Charniak = 43.28, Brown = 41.84 • (no morphology tweaking) • da-da delexicalized: Charniak = 79.62, Brown = 80.20 (!) • (hybrid sv-da Hajič-like tagset = “words”, Penn POS = “tags”) • sv-sv delexicalized: Charniak = 76.07, Brown = 77.01

  19. Unlabeled F Scores Delexicalized Parsing 6/22 • da-da lexicalized: Charniak = 78.16, Brown = 78.24 • (CoNLL train 94K words, test 5852 words) • sv-sv lexicalized: Charniak = 77.81, Brown = 78.74 • (CoNLL train 191K words, test 5656 words) • da-sv lexicalized: Charniak = 43.28, Brown = 41.84 • (no morphology tweaking) • da-da delexicalized: Charniak = 79.62, Brown = 80.20 (!) • (hybrid sv-da Hajič-like tagset = “words”, Penn POS = “tags”) • sv-sv delexicalized: Charniak = 76.07, Brown = 77.01 • da-sv delexicalized: Charniak = 65.50, Brown = 66.40

  20. How Big Swedish Treebank Yields Similar Results? Delexicalized Parsing 7/22 Unlabeled F 1 -score

Recommend

CSC 4181 Compiler Construction Parsing 1 1 Outline Top-down v.s. Bottom-up Top-down parsing

CSC 4181 Compiler Construction Parsing 1 1 Outline Top-down v.s. Bottom-up Top-down parsing

CSC 4181 Compiler Construction Parsing 1 1 Outline Top-down v.s. Bottom-up Top-down parsing Bottom-up parsing Recursive-descent Shift-reduce parsers parsing LR(0) parsing LL(1) parsing LR(0) items LL(1) parsing

429 views • 29 slides
Introduction to Bottom-Up Parsing Shift-reduce parsing The LR parsing algorithm

Introduction to Bottom-Up Parsing Shift-reduce parsing The LR parsing algorithm

Outline Review LL parsing Introduction to Bottom-Up Parsing Shift-reduce parsing The LR parsing algorithm Constructing LR parsing tables 2 Top-Down Parsing: Review Top-Down Parsing: Review Top-down parsing expands a

470 views • 13 slides
Top-Down Parsing Slides modified from Louden Book and Dr. Scherger Top Down Parsing A

Top-Down Parsing Slides modified from Louden Book and Dr. Scherger Top Down Parsing A

Compiler Design and Construction Top-Down Parsing Slides modified from Louden Book and Dr. Scherger Top Down Parsing A top-down parsing algorithm parses an input string of tokens by tracing out the steps in a leftmost derivation. Such an

1.12k views • 96 slides
Statistical Dependency Parsing in Korean: From Corpus Generation To Automatic Parsing Workshop on

Statistical Dependency Parsing in Korean: From Corpus Generation To Automatic Parsing Workshop on

Statistical Dependency Parsing in Korean: From Corpus Generation To Automatic Parsing Workshop on Statistical Parsing of Morphologically-Rich Languages 12th International Conference on Parsing Technologies Jinho D. Choi & Martha Palmer

199 views • 19 slides
Basic Parsing Algorithms Chart Parsing Seminar Recent Advances in Parsing Technology WS

Basic Parsing Algorithms Chart Parsing Seminar Recent Advances in Parsing Technology WS

Basic Parsing Algorithms Chart Parsing Seminar Recent Advances in Parsing Technology WS 2011/2012 Anna Schmidt Talk Outline Chart Parsing Basics Chart Parsing Algorithms Earley Algorithm CKY Algorithm Basics

1.12k views • 73 slides
Overview Introduction Lexicalized TAG, Advantages of parsing with LTAG Parsing LTAGs

Overview Introduction Lexicalized TAG, Advantages of parsing with LTAG Parsing LTAGs

Parsing with Lexicalized TAG (1) Extracting and comparing LTAG (2) Presentation by Philip John Gorinski Seminar Recent Advances in Parsing Technology Saarland University, Winter Term 2011/12 (1) Yves Shabes, Aravind K. Joshi, 1990 (2)

514 views • 36 slides
* 07/16/96 Plan for Today Shift-reduce parsing The problem with predictive top down parsing

* 07/16/96 Plan for Today Shift-reduce parsing The problem with predictive top down parsing

* 07/16/96 Plan for Today Shift-reduce parsing The problem with predictive top down parsing CS453 : Shift Reduce Parsing LR parsing: bottom up Unambiguous Grammars LR(0) and SLR Parse Tables Performs a right-most derivation in

316 views • 6 slides
1 Determinism and Parsing The parsing problem is, given a string w and a context-free grammar G ,

1 Determinism and Parsing The parsing problem is, given a string w and a context-free grammar G ,

1 Determinism and Parsing The parsing problem is, given a string w and a context-free grammar G , to decide if w L ( G ), and if so, to construct a derivation or a parse tree for w . Parsing is studied in courses in compilers. To be efficient

519 views • 12 slides
Bottom-up parsing LR parsing Construct parse tree for input from leaves up LR( k ) parsing

Bottom-up parsing LR parsing Construct parse tree for input from leaves up LR( k ) parsing

Bottom-up parsing LR parsing Construct parse tree for input from leaves up LR( k ) parsing reducing a string of tokens to single start symbol L eft-to-right scan of input, R ightmost derivation (inverse of deriving a string of tokens from

413 views • 5 slides
Bottom Up Parsing Also known as Shift-Reduce parsing More powerful than top down Dont

Bottom Up Parsing Also known as Shift-Reduce parsing More powerful than top down Dont

9/26/2012 Bottom Up Parsing Also known as Shift-Reduce parsing More powerful than top down Dont need left factored grammars Can handle left recursion Bottom Up Parsing Attempt to construct parse tree from an input string

654 views • 6 slides
Dependency Parsing & Feature-based Parsing Ling571 Deep Processing Techniques for NLP

Dependency Parsing & Feature-based Parsing Ling571 Deep Processing Techniques for NLP

Dependency Parsing & Feature-based Parsing Ling571 Deep Processing Techniques for NLP February 2, 2015 Roadmap Dependency parsing Graph-based dependency parsing Maximum spanning tree CLE Algorithm Learning

899 views • 87 slides
Statistical Parsing Parsing context-free languages ar ltekin University of Tbingen

Statistical Parsing Parsing context-free languages ar ltekin University of Tbingen

Statistical Parsing Parsing context-free languages ar ltekin University of Tbingen Seminar fr Sprachwissenschaft November 8, 2016 Recap Parsing basics CKY Earley Closing Ingredients of a (natural language) parser .

1.17k views • 51 slides
Generalised Parsing and Combinator Parsing A Happy Marriage? L. Thomas van Binsbergen

Generalised Parsing and Combinator Parsing A Happy Marriage? L. Thomas van Binsbergen

Conventional Parser Combinators Aspects of Generalised Parsing Combinators for Generalised Parsing Generalised Parsing and Combinator Parsing A Happy Marriage? L. Thomas van Binsbergen ltvanbinsbergen@acm.org Royal Holloway, University of

452 views • 30 slides
Robust Incremental Neural Semantic Graph Parsing Jan Buys and Phil Blunsom Dependency Parsing vs

Robust Incremental Neural Semantic Graph Parsing Jan Buys and Phil Blunsom Dependency Parsing vs

Robust Incremental Neural Semantic Graph Parsing Jan Buys and Phil Blunsom Dependency Parsing vs Semantic Parsing Dependency parsing models the syntactic structure between words in a sentence. Dependency Parsing vs Semantic Parsing

637 views • 31 slides
Statistical Parsing Dependency parsing ar ltekin University of Tbingen Seminar fr

Statistical Parsing Dependency parsing ar ltekin University of Tbingen Seminar fr

Statistical Parsing Dependency parsing ar ltekin University of Tbingen Seminar fr Sprachwissenschaft November 2016 Recap/background Dependency grammar Dependency parsing Evaluation Summary Ingredients of a parser

1.44k views • 89 slides
Plan for Today Predictive parsing as a specific subclass of recursive descent parsing

Plan for Today Predictive parsing as a specific subclass of recursive descent parsing

Plan for Today Predictive parsing as a specific subclass of recursive descent parsing complexity comparisons with general parsing might have to left-factor necessary to remove left-recursion (MiniSVG modification as example) Working

346 views • 5 slides
Lectures 7 and 8. Parsing (syntax analysis) Wei Le 2015.9 Bottom Up Parsing Recognize many

Lectures 7 and 8. Parsing (syntax analysis) Wei Le 2015.9 Bottom Up Parsing Recognize many

Lectures 7 and 8. Parsing (syntax analysis) Wei Le 2015.9 Bottom Up Parsing Recognize many programming languages, used in practice More general, more powerful parser: less restrictions on grammars, e.g., can handle left-recursive grammar

680 views • 64 slides
LL1 Parsing a b $ S 1 A 4 B 3 2 a b b b 1. S --> a B

LL1 Parsing a b $ S 1 A 4 B 3 2 a b b b 1. S --> a B

LL1 Parsing a b $ S 1 A 4 B 3 2 a b b b 1. S --> a B Input 2. B --> 3. B --> A S 4. A --> b B Parse Stack Compare Input and Stacktop 1 LL1 Parsing a b $ S 1 A

694 views • 17 slides
Programming Languages: Parsing Onur Tolga S ehito glu Computer Engineering,METU 27 May

Programming Languages: Parsing Onur Tolga S ehito glu Computer Engineering,METU 27 May

Programming Languages:Parsing Programming Languages: Parsing Onur Tolga S ehito glu Computer Engineering,METU 27 May 2009 Programming Languages:Parsing Outline 1 Parsing Top-down Parsing Recursive Descent Parser LL Parsers

215 views • 8 slides
Models of Human Parsing Experimental Data 2 Informatics 2A: Lecture 22 Eye-tracking Reading

Models of Human Parsing Experimental Data 2 Informatics 2A: Lecture 22 Eye-tracking Reading

Human Parsing Human Parsing Experimental Data Experimental Data Bottom-Up Parser Bottom-Up Parser Left Corner Parser Left Corner Parser Human Parsing 1 Cognitive Constraints Garden Paths Dimensions of Parsing Models of Human Parsing

537 views • 9 slides
Top-Down Parsing Top-Down Parsing #1 Extra Credit Question Given this grammar G: E

Top-Down Parsing Top-Down Parsing #1 Extra Credit Question Given this grammar G: E

Top-Down Parsing Top-Down Parsing #1 Extra Credit Question Given this grammar G: E E + T E T T T * int T int T ( E ) Is the string int * (int + int) in L(G)? Give a derivation or prove that it

1.08k views • 61 slides
CKY Parsing Ling 571 Deep Processing Techniques for NLP January 12, 2011 Roadmap

CKY Parsing Ling 571 Deep Processing Techniques for NLP January 12, 2011 Roadmap

CKY Parsing Ling 571 Deep Processing Techniques for NLP January 12, 2011 Roadmap Motivation: Parsing (In) efficiency Dynamic Programming Cocke-Kasami-Younger Parsing Algorithm Chomsky Normal Form Conversion

689 views • 42 slides
Parsing Parsing involves: determining if a string belongs to a language, and

Parsing Parsing involves: determining if a string belongs to a language, and

Parsing Parsing involves: determining if a string belongs to a language, and constructing structure of string if it belongs to language. Two approaches to constructing parsers: 1. Top down parsing: Our focus is on table-driven

428 views • 18 slides
Compiler Construction Lecture 7: Bottom-up parsing 2020-01-28 Michael Engel Includes material

Compiler Construction Lecture 7: Bottom-up parsing 2020-01-28 Michael Engel Includes material

Compiler Construction Lecture 7: Bottom-up parsing 2020-01-28 Michael Engel Includes material by Jan Christian Meyer and Rich Maclin (UNM) Overview Top-down parsing revisited Bottom-up parsing Comparison to top-down parsing

539 views • 20 slides

More recommend