1 / 50

Semi-Supervised Approaches for Learning to Parse Natural Languages

Semi-Supervised Approaches for Learning to Parse Natural Languages. Slides are from Rebecca Hwa, Ray Mooney. The Role of Parsing in Language Applications…. As a stand-alone application Grammar checking As a pre-processing step Question Answering Information extraction

kalona
Download Presentation

Semi-Supervised Approaches for Learning to Parse Natural Languages

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. Semi-Supervised Approaches for Learning to Parse Natural Languages Slides are from Rebecca Hwa, Ray Mooney

  2. The Role of Parsing in Language Applications… • As a stand-alone application • Grammar checking • As a pre-processing step • Question Answering • Information extraction • As an integral part of a model • Speech Recognition • Machine Translation

  3. Parsing Input: I saw her S • Parsers provide syntactic analyses of sentences NP VP PN VB NP I saw PN her

  4. Challenges in Building Parsers • Disambiguation • Lexical disambiguation • Structural disambiguation • Rule Exceptions • Many lexical dependencies • Manual Grammar Construction • Limited coverage • Difficult to maintain

  5. Meeting these Challenges:Statistical Parsing • Disambiguation? • Resolve local ambiguities with global likelihood • Rule Exceptions? • Lexicalized representation • Manual Grammar Construction? • Automatic induction from large corpora • A new challenge: how to obtain training corpora? • Make better use of unlabeled data with machine learning techniquesand linguistic knowledge

  6. Roadmap • Parsing as a learning problem • Semi-supervised approaches • Sample selection • Co-training • Corrected Co-training • Conclusion and further directions

  7. Parsing Ambiguities Input: “I saw her duck with a telescope” T1: S T2: S VP NP VP NP PN VB NP PP PN NP VB PNS N P NP PNS N PP I saw I saw her duck with DET N her duck P NP with DET N a telescope a telescope

  8. Disambiguation with Statistical Parsing W = “I saw her duck with a telescope” T1: S T2: S VP NP VP NP PN VB NP PP PN NP VB PNS N P NP PNS N PP I saw I saw her duck with DET N her duck P NP with DET N a telescope a telescope

  9. A Statistical Parsing Model Example of PCFG rules: DET N 0.7 NP • Probabilistic Context-Free Grammar (PCFG) • Associate probabilities with production rules • Likelihood of the parse is computed from the rules used • Learn rule probabilities from training data 0.3 NP PN a 0.5DET an 0.1DET the 0.4 DET ...

  10. Sentence Probability • Assume productions for each node are chosen independently. • Probability of derivation is the product of the probabilities of its productions. P(D1) = 0.1 x 0.5 x 0.5 x 0.6 x 0.6 x 0.5 x 0.3 x 1.0 x 0.2 x 0.2 x 0.5 x 0.8 = 0.0000216 D1 S 0.1 VP 0.5 Verb NP 0.6 0.5 Det Nominal book 0.5 0.6 Nominal PP the 1.0 0.3 Prep NP Noun 0.2 0.2 0.5 Proper-Noun flight through 0.8 Houston

  11. Syntactic Disambiguation • Resolve ambiguity by picking most probable parse tree. S D2 P(D2) = 0.1 x 0.3 x 0.5 x 0.6 x 0.5 x 0.6 x 0.3 x 1.0 x 0.5 x 0.2 x 0.2 x 0.8 = 0.00001296 0.1 VP 0.3 VP 0.5 Verb NP 0.6 0.5 PP Det Nominal book 1.0 0.6 0.3 Noun Prep NP the 0.2 0.2 0.5 Proper-Noun flight through 0.8 Houston 11

  12. Sentence Probability • Probability of a sentence is the sum of the probabilities of all of its derivations. P(“book the flight through Houston”) = P(D1) + P(D2) = 0.0000216 + 0.00001296 = 0.00003456

  13. Supervised PCFG Training S →NP VP S → VP NP → Det A N NP → NP PP NP → PropN A → ε A → Adj A PP → Prep NP VP → V NP VP → VP PP 0.9 0.1 0.5 0.3 0.2 0.6 0.4 1.0 0.7 0.3 S S NP VP NP VP John V NP PP John V NP PP put the dog in the pen put the dog in the pen English PCFG: Supervised Training • If parse trees are provided for training sentences, a grammar and its parameters can be can all be estimated directly from counts accumulated from the tree-bank (with appropriate smoothing). Tree Bank . . .

  14. Estimating Production Probabilities • Set of production rules can be taken directly from the set of rewrites in the treebank. • Parameters can be directly estimated from frequency counts in the treebank.

  15. Vanilla PCFG Limitations • Since probabilities of productions do not rely on specific words or concepts, only general structural disambiguation is possible (e.g. prefer to attach PPs to Nominals). • Consequently, vanilla PCFGs cannot resolve syntactic ambiguities that require semantics to resolve, e.g. ate with fork vs. meatballs. • In order to work well, PCFGs must be lexicalized, i.e. productions must be specialized to specific words by including their head-word in their LHS non-terminals (e.g. VP-ate).

  16. John put the dog in the pen. S S →NP VP S → VP NP → Det A N NP → NP PP NP → PropN A → ε A → Adj A PP → Prep NP VP → V NP VP → VP PP 0.9 0.1 0.5 0.3 0.2 0.6 0.4 1.0 0.7 0.3 NP VP PCFG Parser John V NP PP put the dog in the pen English Example of Importance of Lexicalization • A general preference for attaching PPs to NPs rather than VPs can be learned by a vanilla PCFG. • But the desired preference can depend on specific words. 16

  17. Example of Importance of Lexicalization A general preference for attaching PPs to NPs rather than VPs can be learned by a vanilla PCFG. But the desired preference can depend on specific words. John put the dog in the pen. S →NP VP S → VP NP → Det A N NP → NP PP NP → PropN A → ε A → Adj A PP → Prep NP VP → V NP VP → VP PP 0.9 0.1 0.5 0.3 0.2 0.6 0.4 1.0 0.7 0.3 PCFG Parser English S X NP VP John V NP put the dog in the pen 17

  18. Head Words Syntactic phrases usually have a word in them that is most “central” to the phrase. Linguists have defined the concept of a lexical head of a phrase. Simple rules can identify the head of any phrase by percolating head words up the parse tree. Head of a VP is the main verb Head of an NP is the main noun Head of a PP is the preposition Head of a sentence is the head of its VP

  19. Lexicalized Productions Specialized productions can be generated by including the head word and its POS of each non-terminal as part of that non-terminal’s symbol. S liked-VBD liked-VBD NP VP John-NNP dog-NN VBD NP NNP Nominaldog-NN → Nominaldog-NN PPin-IN DT Nominal liked John dog-NN Nominal PP the in-IN dog-NN IN NP NN pen-NN dog in DT Nominal pen-NN the NN pen

  20. Lexicalized Productions S put-VBD NP VP VPput-VBD → VPput-VBD PPin-IN put-VBD John-NNP VP PP NNP in-IN put-VBD John IN NP VBD NP pen-NN dog-NN put in DT Nominal pen-NN DT Nominal dog-NN the NN the NN pen dog

  21. Parameterizing Lexicalized Productions Accurately estimating parameters on such a large number of very specialized productions could require enormous amounts of treebank data. Need some way of estimating parameters for lexicalized productions that makes reasonable independence assumptions so that accurate probabilities for very specific rules can be learned.

  22. Collins’ Parser Collins’ (1999) parser assumes a simple generative model of lexicalized productions. Models productions based on context to the left and the right of the head daughter. LHS → LnLn1…L1H R1…Rm1Rm First generate the head (H) and then repeatedly generate left (Li) and right (Ri) context symbols until the symbol STOP is generated.

  23. Sample Production Generation VPput-VBD → VBDput-VBD NPdog-NN PPin-IN VPput-VBD → STOP VBDput-VBD NPdog-NN PPin-IN STOP L1 H R1 R2 R3 PL(STOP | VPput-VBD) * PH(VBD | VPput-VBD)* PR(NPdog-NN | VPput-VBD)* PR(PPin-IN | VPput-VBD) * PR(STOP | VPput-VBD)

  24. Estimating Production Generation Parameters Estimate PH, PL, and PR parameters from treebank data. Count(PPin-IN right of head in a VPput-VBD production) PR(PPin-IN| VPput-VBD) = Count(symbol right of head in a VPput-VBD) Count(NPdog-NN right of head in a VPput-VBD production) PR(NPdog-NN| VPput-VBD) = Count(symbol right of head in a VPput-VBD) • Smooth estimates by linearly interpolating with simpler models conditioned on just POS tag or no lexical info. smPR(PPin-IN| VPput-VBD) = 1 PR(PPin-IN| VPput-VBD) + (11) (2 PR(PPin-IN| VPVBD) + (12) PR(PPin-IN| VP))

  25. Parsing Evaluation Metrics • PARSEVAL metrics measure the fraction of the constituents that match between the computed and human parse trees. If P is the system’s parse tree and T is the human parse tree (the “gold standard”): • Recall= (# correct constituents in P) / (# constituents in T) • Precision = (# correct constituents in P) / (# constituents in P) • Labeled Precisionand labeled recallrequire getting the non-terminal label on the constituent node correct to count as correct. • F1 is the harmonic mean of precision and recall.

  26. Treebank Results • Results of current state-of-the-art systems on the English Penn WSJ treebank are slightly greater than 90% labeled precision and recall.

  27. Supervised Learning Avoids Manual Construction • Training examples are pairs of problems and answers • Training examples for parsing: a collection of sentence, parse tree pairs (Treebank) • From the treebank, get maximum likelihood estimates for the parsing model • New challenge: treebanks are difficult to obtain • Needs human experts • Takes years to complete

  28. Hwa’s Approach • Sample selection • Reduce the amount of training data by picking more useful examples • Co-training • Improve parsing performance from unlabeled data • Corrected Co-training • Combine ideas from both sample selection and co-training

  29. Roadmap • Parsing as a learning problem • Semi-supervised approaches • Sample selection • Overview • Scoring functions • Evaluation • Co-training • Corrected Co-training • Conclusion and further directions

  30. Sample Selection • Assumption • Have lots of unlabeled data (cheap resource) • Have a human annotator (expensive resource) • Iterative training session • Learner selects sentences to learn from • Annotator labels these sentences • Goal: Predict the benefit of annotation • Learner selects sentences with the highest Training Utility Values (TUVs) • Key issue: scoring function to estimate TUV

  31. Algorithm Initialize Train the parser on a small treebank (seed data) to get the initial parameter values. Repeat Create candidate set by randomly sample the unlabeled pool. Estimate the TUV of each sentence in the candidate set with a scoring function, f. Pick the n sentences with the highest score (according to f). Human labels these n sentences and add them to training set. Re-train the parser with the updated training set. Until (no more data).

  32. Scoring Function • Approximate the TUV of each sentence • True TUVs are not known • Need relative ranking • Ranking criteria • Knowledge about the domain • e.g., sentence clusters, sentence length, … • Output of the hypothesis • e.g., error-rate of the parse, uncertainty of the parse, … ….

  33. f f f error te len Proposed Scoring Functions • Using domain knowledge • long sentences tend to be complex • Uncertainty about the output of the parser • tree entropy • Minimize mistakes made by the parser • use an oracle scoring function find sentences with the most parsing inaccuracies

  34. Entropy • Measure of uncertainty in a distribution • Uniform distribution very uncertain • Spike distribution very certain • Expected number of bits for encoding a probability distribution, X

  35. Tree Entropy Scoring Function • Distribution over parse trees for sentence W: • Tree entropy: uncertainty of the parse distribution • Scoring function: ratio of actual parse tree entropy to that of a uniform distribution

  36. Experimental Setup • Parsing model: • Collins Model 2 • Candidate pool • WSJ sec 02-21, with the annotation stripped • Initial labeled examples: 500 sentences • Per iteration: add 100 sentences • Testing metric: f-score (precision/recall) • Test data: • ~2000 unseen sentences (from WSJ sec 00) • Baseline • Annotate data in sequential order

  37. Training Examples Vs. Parsing Performance

  38. Parsing Performance Vs. Constituents Labeled

  39. Co-Training [Blum and Mitchell, 1998] • Assumptions • Have a small treebank • No further human assistance • Have two different kinds of parsers • A subset of each parser’s output becomes new training data for the other • Goal: • select sentences that are labeled with confidence by one parser but labeled with uncertainty by the other parser.

  40. Algorithm Initialize Train two parsers on a small treebank (seed data) to get the initial models. Repeat Create candidate set by randomly sample the unlabeled pool. Each parser labels the candidate set and estimates the accuracy of its output with scoring function, f. Choose examples according to some selection method, S (using the scores from f). Add them to the parsers’ training sets. Re-train parsers with the updated training sets. Until (no more data).

  41. Scoring Functions • Evaluates the quality of each parser’s output • Ideally, function measures accuracy • Oracle fF-score • combined prec./rec. of the parse • Practical scoring functions • Conditional probability fcprob • Prob(parse | sentence) • Others (joint probability, entropy, etc.)

  42. Selection Methods • Above-n: Sabove-n • The score of the teacher’s parse is greater than n • Difference: Sdiff-n • The score of the teacher’s parse is greater than that of the student’s parse by n • Intersection: Sint-n • The score of the teacher’s parse is one of its n% highest while the score of the student’s parse for the same sentence is one of the student’s n% lowest

  43. Experimental Setup • Co-training parsers: • Lexicalized Tree Adjoining Grammar parser [Sarkar, 2002] • Lexicalized Context Free Grammar parser [Collins, 1997] • Seed data: 1000 parsed sentences from WSJ sec02 • Unlabeled pool: rest of the WSJ sec02-21, stripped • Consider 500 unlabeled sentences per iteration • Development set: WSJ sec00 • Test set: WSJ sec23 • Results: graphs for the Collins parser

  44. Co-Training using fcprob

  45. Roadmap • Parsing as a learning problem • Semi-supervised approaches • Sample selection • Co-training • Corrected Co-training • Conclusion and further directions

  46. Corrected Co-Training • Human reviews and corrects the machine outputs before they are added to the training set • Can be seen as a variant of sample selection [cf. Muslea et al., 2000] • Applied to Base NP detection [Pierce & Cardie, 2001]

  47. Algorithm Initialize: Train two parsers on a small treebank (seed data) to get the initial models. Repeat Create candidate set by randomly sample the unlabeled pool. Each parser labels the candidate set and estimates the accuracy of its output with scoring function, f. Choose examples according to some selection method, S (using the scores from f). Human reviews and corrects the chosen examples. Add them to the parsers’ training sets. Re-train parsers with the updated training sets. Until (no more data).

  48. Selection Methods and Corrected Co-Training • Two scoring functions: fF-score , fcprob • Three selection methods: Sabove-n , Sdiff-n , Sint-n

  49. Corrected Co-Training using fcprob (Reviews)

More Related