Transition-based Dependency Parsing
with Selectional Branching
Presented at the 4th workshop on
Statistical Parsing in Mor...
Greedy vs. Non-greedy Parsing

•

•

Greedy parsing

-

Considers only one head for each token.
Generates one parse tree p...
Motivation

•

How often do we need non-greedy parsing?

•

Our greedy parser performs as accurately as our non-greedy
par...
Transition-based Parsing

•

Transition-based dependency parsing (greedy)

-

Considers one transition for each parsing st...
Transition-based Parsing

•

Transition-based dependency parsing (greedy)

-

Considers one transition for each parsing st...
Transition-based Parsing

•

Transition-based dependency parsing (greedy)

-

t1

…

S

Considers one transition for each ...
Transition-based Parsing

•

Transition-based dependency parsing (greedy)

-

t1

t′

…

S

Considers one transition for e...
Transition-based Parsing

•

Transition-based dependency parsing (greedy)

-

t1

t′

S

…

S

Considers one transition fo...
Transition-based Parsing
Transition-based dependency parsing (greedy)

-

t1

t′

S

t1

…

S

Considers one transition fo...
Transition-based Parsing
Transition-based dependency parsing (greedy)

-

t1

t′

S

t1

…

S

Considers one transition fo...
Transition-based Parsing
Transition-based dependency parsing (greedy)

-

t1

t′

S

t1

…

S

Considers one transition fo...
Transition-based Parsing
Transition-based dependency parsing (greedy)

-

t1

t′

S

t1

…

S

Considers one transition fo...
Transition-based Parsing
Transition-based dependency parsing (greedy)

-

t1

t′

S

t1

…

S

Considers one transition fo...
Transition-based Parsing

•

Transition-based dependency parsing with beam search

-

Considers b-num. of transitions for ...
Transition-based Parsing

•

Transition-based dependency parsing with beam search

-

Considers b-num. of transitions for ...
Transition-based Parsing

•

Transition-based dependency parsing with beam search

-

t1

…

S

Considers b-num. of transi...
Transition-based Parsing

•

Transition-based dependency parsing with beam search

-

t1

t′1

…

…

S

Considers b-num. o...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

t′b

t1

Sb

…

S

Considers b-num...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

t1L

t′b

t1

Sb

tb1

t11

…

…

...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

t1L

t′b

t1

Sb

tb1

t11

…

…

...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

t1L

t′b

t1

Sb

tb1

t11

t′1

…...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

Sb

tb1

t′1 …

S1

t′b …

Sb

t1L...
Transition-based Parsing
Transition-based dependency parsing with beam search

t′1

S1

Sb

tb1

t′1 …

S1

t′b …

Sb

t1L...
Selectional Branching

•

Issues with beam search
Generates the fixed number of parse trees no matter how
easy/hard the inp...
Selectional Branching
S1

7
Wednesday, October 23, 13
Selectional Branching
t11

…

S1

t1L

7
Wednesday, October 23, 13
Selectional Branching
t11

t′11

…

S1

t1L

7
Wednesday, October 23, 13
Selectional Branching
t11

…

S1

t1L

t′11

low
confident?

7
Wednesday, October 23, 13
Selectional Branching
t11

…

S1

t1L

t′11

low
confident?

λ=

7
Wednesday, October 23, 13
Selectional Branching
t11

…

S1

t1L

λ=

S1

t′11

low
confident?
t′12 …

S1

t′1k

7
Wednesday, October 23, 13
Selectional Branching
t11

…

S1

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t′1k

7
Wednesday, October 23, 13
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

t′1k

7
Wednesday, Octobe...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

t′1k

7
Wednesday, Octobe...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

t′1k

7
Wednesday, Octobe...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

S2

t′1k

7
Wednesday, Oc...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

S2

t′1k

7
Wednesday, Oc...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

S2

t′1k

7
Wednesday, Oc...
Selectional Branching

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

t11

…

S1

t2L

S2

t′1k

7
Wednesday, Oc...
Selectional Branching

…

t11

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

S1

t2L

S2

t′1k

t′21 …

Sn

low...
Selectional Branching

…

t11

t1L

λ=

S1

S2

t′11

low
confident?
t′12 …

S1

t21

…

S1

t2L

S2

t′1k

t′21 …

Sn

low...
Selectional Branching

8
Wednesday, October 23, 13
Selectional Branching
λ=

S1

t′12

S2

8
Wednesday, October 23, 13

t′22

S3

t′32
Selectional Branching
λ=
S1

t′12

S1

t′12

S2

S2 …

Sa

8
Wednesday, October 23, 13

t′22

S3

t′32

T
Selectional Branching
λ=
S1

S1

S2 …

t′12
S2

S2

t′12

t′22

Sa

S3 …

8
Wednesday, October 23, 13

S3

t′22

t′32

T
S...
Selectional Branching
λ=
S1

S1

S3

Sa

S3 …

t′22

S4 …

t′32

8
Wednesday, October 23, 13

S3

t′22

S2 …

t′12
S2

S2
...
Selectional Branching
λ=
S1

S1

S3

Sa

S3 …

t′22

S3

t′22

S2 …

t′12
S2

S2

t′12

S4 …

t′32

t′32

T
Sb

T
Sc

T

C...
Selectional Branching
λ=
S1

S1

S3

Sa

S3 …

t′22

S3

t′22

S2 …

t′12
S2

S2

t′12

S4 …

t′32

t′32

T
Sb

T
Sc

T

C...
…

…

a classifier C 1 that uses a feature map (x, y) and
a weight1vector w to measure a score for each label
…
ssdt
2t
C (...
Experiments

•

•

Parsing algorithm (Choi & McCallum, 2013)

-

Hybrid between Nivre’s arc-eager and list-based algorithm...
ber of A DAG RAD iterations. Using an Intel Xeon
The thir
2.57GHz machine, it takes less than 40 minutes
external
to train...
Projective Parsing

•

The benchmark setup using WSJ.
Approach

USA

LAS

Time

bt = 80, bd = 80

92.96

91.93

0.009

bt ...
Projective Parsing

•

The benchmark setup using WSJ.
Approach

LAS

Time

bt = 80, bd = 80

92.96

91.93

0.009

Zhang & ...
Non-projective Parsing

•

CoNLL-X shared task data
Approach

Danish

Dutch

Slovene

Swedish

LAS UAS LAS UAS LAS UAS LAS...
SPMRL 2013 Shared Task

•

Baseline results provided by ClearNLP.
Language

5K

Full

LAS

UAS

LS

LAS

UAS

LS

Arabic

...
Conclusion

•

•

Selectional branching

-

Uses confidence estimates to decide when to employ a beam.
Shows comparable acc...
Upcoming SlideShare
Loading in …5
×

Transition-based Dependency Parsing with Selectional Branching

852 views

Published on

We present a novel approach, called selectional branching, which uses confidence estimates to decide when to employ a beam, providing the accuracy of beam search at speeds close to a greedy transition-based dependency parsing approach. Selectional branching is guaranteed to perform a fewer number of transitions than beam search yet performs as accurately. We also present a new transition-based dependency parsing algorithm that gives a complexity of O(n) for projective parsing and an expected linear time speed for non-projective parsing. With the standard setup, our parser shows an unlabeled attachment score of 92.96% and a parsing speed of 9 milliseconds per sentence, which is faster and more accurate than the current state-of-the-art transition- based parser that uses beam search.

Published in: Technology, Business
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
852
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
9
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Transition-based Dependency Parsing with Selectional Branching

  1. 1. Transition-based Dependency Parsing with Selectional Branching Presented at the 4th workshop on Statistical Parsing in Morphologically Rich Languages October 18th, 2013 Jinho D. Choi University of Massachusetts Amherst Wednesday, October 23, 13
  2. 2. Greedy vs. Non-greedy Parsing • • Greedy parsing - Considers only one head for each token. Generates one parse tree per sentence. e.g., transition-based parsing (2 ms / sentence). Non-greedy parsing - Considers multiple heads for each token. Generates multiple parse trees per sentence. e.g., transition-based parsing with beam search, graph-based parsing, linear programming, dual decomposition (≥ 93%). 2 Wednesday, October 23, 13
  3. 3. Motivation • How often do we need non-greedy parsing? • Our greedy parser performs as accurately as our non-greedy parser about 64% of the time. - • - This gap is even closer when they are evaluated on nonbenchmark data (e.g., twits, chats, blogs). Many applications are time sensitive. - Some applications need at least one complete parse tree ready given a limited time period (e.g., search, dialog, Q/A). Hard sentences are hard for any parser! - Considering more heads does not always guarantee more accurate parse results. 3 Wednesday, October 23, 13
  4. 4. Transition-based Parsing • Transition-based dependency parsing (greedy) - Considers one transition for each parsing state. 4 Wednesday, October 23, 13
  5. 5. Transition-based Parsing • Transition-based dependency parsing (greedy) - Considers one transition for each parsing state. S 4 Wednesday, October 23, 13
  6. 6. Transition-based Parsing • Transition-based dependency parsing (greedy) - t1 … S Considers one transition for each parsing state. tL 4 Wednesday, October 23, 13
  7. 7. Transition-based Parsing • Transition-based dependency parsing (greedy) - t1 t′ … S Considers one transition for each parsing state. tL 4 Wednesday, October 23, 13
  8. 8. Transition-based Parsing • Transition-based dependency parsing (greedy) - t1 t′ S … S Considers one transition for each parsing state. tL 4 Wednesday, October 23, 13
  9. 9. Transition-based Parsing Transition-based dependency parsing (greedy) - t1 t′ S t1 … S Considers one transition for each parsing state. … • tL tL 4 Wednesday, October 23, 13
  10. 10. Transition-based Parsing Transition-based dependency parsing (greedy) - t1 t′ S t1 … S Considers one transition for each parsing state. … • tL tL 4 Wednesday, October 23, 13 t′
  11. 11. Transition-based Parsing Transition-based dependency parsing (greedy) - t1 t′ S t1 … S Considers one transition for each parsing state. … • tL tL 4 Wednesday, October 23, 13 t′ … S
  12. 12. Transition-based Parsing Transition-based dependency parsing (greedy) - t1 t′ S t1 … S Considers one transition for each parsing state. … • tL tL 4 Wednesday, October 23, 13 t′ … S T
  13. 13. Transition-based Parsing Transition-based dependency parsing (greedy) - t1 t′ S t1 … S Considers one transition for each parsing state. … • tL tL 4 Wednesday, October 23, 13 t′ … S T What if t′ is not the correct transition?
  14. 14. Transition-based Parsing • Transition-based dependency parsing with beam search - Considers b-num. of transitions for each block of parsing states. 5 Wednesday, October 23, 13
  15. 15. Transition-based Parsing • Transition-based dependency parsing with beam search - Considers b-num. of transitions for each block of parsing states. S 5 Wednesday, October 23, 13
  16. 16. Transition-based Parsing • Transition-based dependency parsing with beam search - t1 … S Considers b-num. of transitions for each block of parsing states. tL 5 Wednesday, October 23, 13
  17. 17. Transition-based Parsing • Transition-based dependency parsing with beam search - t1 t′1 … … S Considers b-num. of transitions for each block of parsing states. t′b tL 5 Wednesday, October 23, 13
  18. 18. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 t′b t1 Sb … S Considers b-num. of transitions for each block of parsing states. … - … • tL 5 Wednesday, October 23, 13
  19. 19. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 t1L t′b t1 Sb tb1 t11 … … … S Considers b-num. of transitions for each block of parsing states. … - … • tL tbL 5 Wednesday, October 23, 13
  20. 20. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 t1L t′b t1 Sb tb1 t11 … … … S Considers b-num. of transitions for each block of parsing states. … - … • tL tbL 5 Wednesday, October 23, 13
  21. 21. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 t1L t′b t1 Sb tb1 t11 t′1 … … … S Considers b-num. of transitions for each block of parsing states. … - … • tL tbL 5 Wednesday, October 23, 13 t′b
  22. 22. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 Sb tb1 t′1 … S1 t′b … Sb t1L t′b t1 t11 … … … S Considers b-num. of transitions for each block of parsing states. … - … • tL tbL 5 Wednesday, October 23, 13
  23. 23. Transition-based Parsing Transition-based dependency parsing with beam search t′1 S1 Sb tb1 t′1 … S1 t′b … Sb t1L t′b t1 t11 T1 … … … S Considers b-num. of transitions for each block of parsing states. … - … • tL tbL 5 Wednesday, October 23, 13 Tb
  24. 24. Selectional Branching • Issues with beam search Generates the fixed number of parse trees no matter how easy/hard the input sentence is. - • - Is it possible to dynamically adjust the beam size for each individual sentence? Selectional branching - One-best transition sequence is found by a greedy parser. - Generate transition sequences from the b-1 highest scoring state-transition pairs in the collection. Collect k-best state-transition pairs for each low confidence transition used to generate the one-best sequence. 6 Wednesday, October 23, 13
  25. 25. Selectional Branching S1 7 Wednesday, October 23, 13
  26. 26. Selectional Branching t11 … S1 t1L 7 Wednesday, October 23, 13
  27. 27. Selectional Branching t11 t′11 … S1 t1L 7 Wednesday, October 23, 13
  28. 28. Selectional Branching t11 … S1 t1L t′11 low confident? 7 Wednesday, October 23, 13
  29. 29. Selectional Branching t11 … S1 t1L t′11 low confident? λ= 7 Wednesday, October 23, 13
  30. 30. Selectional Branching t11 … S1 t1L λ= S1 t′11 low confident? t′12 … S1 t′1k 7 Wednesday, October 23, 13
  31. 31. Selectional Branching t11 … S1 t1L λ= S1 S2 t′11 low confident? t′12 … S1 t′1k 7 Wednesday, October 23, 13
  32. 32. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L t′1k 7 Wednesday, October 23, 13
  33. 33. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L t′1k 7 Wednesday, October 23, 13 t′21
  34. 34. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L t′1k 7 Wednesday, October 23, 13 t′21 low confident?
  35. 35. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L S2 t′1k 7 Wednesday, October 23, 13 t′21 low confident? t′22 … S2 t′2k
  36. 36. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L S2 t′1k 7 Wednesday, October 23, 13 t′21 … Sn low confident? t′22 … S2 t′2k
  37. 37. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L S2 t′1k 7 Wednesday, October 23, 13 t′21 … Sn low confident? t′22 … S2 t′2k …
  38. 38. Selectional Branching t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … t11 … S1 t2L S2 t′1k 7 Wednesday, October 23, 13 t′21 … Sn low confident? t′22 … S2 t′2k … T
  39. 39. Selectional Branching … t11 t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … S1 t2L S2 t′1k t′21 … Sn low confident? t′22 … S2 t′2k … Pick b-1 number of pairs with the highest scores. 7 Wednesday, October 23, 13 T
  40. 40. Selectional Branching … t11 t1L λ= S1 S2 t′11 low confident? t′12 … S1 t21 … S1 t2L S2 t′1k t′21 … Sn low confident? t′22 … S2 t′2k … Pick b-1 number of pairs with the highest scores. For our experiments, k = 2 is used. 7 Wednesday, October 23, 13 T
  41. 41. Selectional Branching 8 Wednesday, October 23, 13
  42. 42. Selectional Branching λ= S1 t′12 S2 8 Wednesday, October 23, 13 t′22 S3 t′32
  43. 43. Selectional Branching λ= S1 t′12 S1 t′12 S2 S2 … Sa 8 Wednesday, October 23, 13 t′22 S3 t′32 T
  44. 44. Selectional Branching λ= S1 S1 S2 … t′12 S2 S2 t′12 t′22 Sa S3 … 8 Wednesday, October 23, 13 S3 t′22 t′32 T Sb T
  45. 45. Selectional Branching λ= S1 S1 S3 Sa S3 … t′22 S4 … t′32 8 Wednesday, October 23, 13 S3 t′22 S2 … t′12 S2 S2 t′12 t′32 T Sb T Sc T
  46. 46. Selectional Branching λ= S1 S1 S3 Sa S3 … t′22 S3 t′22 S2 … t′12 S2 S2 t′12 S4 … t′32 t′32 T Sb T Sc T Carries on parsing states from the one-best sequence. 8 Wednesday, October 23, 13
  47. 47. Selectional Branching λ= S1 S1 S3 Sa S3 … t′22 S3 t′22 S2 … t′12 S2 S2 t′12 S4 … t′32 t′32 T Sb T Sc T Carries on parsing states from the one-best sequence. Guarantees to generate fewer trees than beam search when |λ| ≤ b. 8 Wednesday, October 23, 13
  48. 48. … … a classifier C 1 that uses a feature map (x, y) and a weight1vector w to measure a score for each label … ssdt 2t C (x) = arg max{f (x, y)} … y 2 Y, and choosing a label with the highest score. y2Y When there is a tie between labels(x, y)) highest exp(w · with the … s3t f (x, y) = P ranching strategy. 1 score, the first one is chosen. This canhighest scoring be y 0 )) Let C be a classifier that2Y exp(w · (x, expressed 0 finds the y om T1 . as logistic regression: transitiona given the parsing state x. To find low confidence predictions, we use the marC 1 (x) = arg max{f (x, y)} prediction T1 =…[s11 , ... , ss1t ] dt gins (score differences) between the best y2Y While generating and the other predictions. If all ·margins are greater exp(w (x, y)) p2j ), ... , (s1j , pkj ) than a threshold, the best prediction is considered f (x, y) = P ranching strategy. (x, y 0 )) 0 2Y exp(w · y ce T . highly confident; otherwise, it is not. Given this om prediction p1j 1 sequences are gen- k analogy, the k-best predictions can be found as Let C To findclassifier that finds the k-highest scoring be a low confidence predictions, we use the maring predictions in follows (m 0 is a margin threshold): prediction T1 = [s11 , ...transitions(score differences) between the x and the margin m. , s1t ] gins given the parsing state best | < b, all predicWhile generating and k other predictions. If all margins are greater the edy ... , (s is pkj ) C (x, m) = K arg max{f (x, y)} p2j ),parser 1j ,used than a threshold, the best prediction is considered y2Y sce predictionnow although it p1j highly confident; otherwise, it is not. Given m s.t. f (x, C 1 (x)) f (x, y)  this tial parsing state, sequences are gen- analogy, the k-best predictions can be found as further transitions. ing predictions in follows (m 0 transition labels The highest max’ returns margin threshold):whose mar‘K arg scoringis a a set of k 0C1(x) is low confident if generated, a parse | < b, all predic|Ck(x, m)| kto1. 1 (x) are smaller than any other label’s the parser isscore. gins> C 1 = K arg max{f (x, y)} 0 edy highest used C (x, m) margin to C (x) and also  m, where k  k. y2Y = 2, which gave s although it now it returns aC 1 of the f (x, y)  m stial parsing1. We than k = state, When m = 0, s.t. f (x, 9set(x)) highest scoring labels only, including C 1 (x). When m = 1, it rehich did not23, 13 show further transitions. Wednesday, October 0 Low Confidence Transition • • … … •
  49. 49. Experiments • • Parsing algorithm (Choi & McCallum, 2013) - Hybrid between Nivre’s arc-eager and list-based algorithms. Projective parsing: O(n). Non-projective parsing: expected linear time. Features - Rich non-local features from Zhang & Nivre, 2011. - For languages with morphological features, morphologies of σ[0] and β[0] are used as unigram features. For languages with coarse-grained POS tags, feature templates using fine-grained POS tags are replicated. 10 Wednesday, October 23, 13
  50. 50. ber of A DAG RAD iterations. Using an Intel Xeon The thir 2.57GHz machine, it takes less than 40 minutes external to train the entire Penn Treebank, which includes our appr #times for IO, feature extraction and bootstrapping.sizes. of transitions performed with respect to beam seconds Number of Transitions • 1,200,000 Transitions 1,000,000 800,000 600,000 400,000 200,000 0 0 10 20 30 40 50 60 70 80 Beam size = 1, 2, 4, 8, 16, 32, 64, 80 Figure 5: The total number of transitions performed 11 during decoding with respect to beam sizes on the Wednesday, October 23, 13 Approa Zhang a Huang a Zhang a Bohnet a McDona Mcdona Sagae an Koo and Zhang a Martins Rush et Koo et a Carreras Bohnet a Suzuki e bt = 80
  51. 51. Projective Parsing • The benchmark setup using WSJ. Approach USA LAS Time bt = 80, bd = 80 92.96 91.93 0.009 bt = 80, bd = 64 92.96 91.93 0.009 bt = 80, bd = 32 92.96 91.94 0.009 bt = 80, bd = 16 92.96 91.94 0.008 bt = 80, bd = 8 92.89 91.87 0.006 bt = 80, bd = 4 92.76 91.76 0.004 bt = 80, bd = 2 92.56 91.54 0.003 bt = 80, bd = 1 92.26 91.25 0.002 bt = 1, bd = 1 92.06 91.05 0.002 12 Wednesday, October 23, 13
  52. 52. Projective Parsing • The benchmark setup using WSJ. Approach LAS Time bt = 80, bd = 80 92.96 91.93 0.009 Zhang & Clark, 2008 92.1 Huang & Sagae, 2010 92.1 Zhang & Nivre, 2011 92.9 91.8 0.03 Bohnet & Nivre, 2012 93.38 92.44 0.4 McDonald et al., 2005 90.9 McDonald & Pereira, 2006 91.5 Sagae & Lavie, 2006 92.7 Koo & Collins, 2010 93.04 Zhang & McDonald, 2012 93.06 Martins et al., 2010 93.26 Rush et al., 2010 Wednesday, October 23, 13 USA 93.8 13 0.04 91.86
  53. 53. Non-projective Parsing • CoNLL-X shared task data Approach Danish Dutch Slovene Swedish LAS UAS LAS UAS LAS UAS LAS UAS bt = 80, bd = 80 87.27 91.36 82.45 85.33 77.46 84.65 86.80 91.36 bt = 80, bd = 1 86.75 91.04 80.75 83.59 75.66 83.29 86.32 91.12 Nivre et al., 2006 84.77 89.80 78.59 81.35 70.30 78.72 84.58 89.50 McDonald et al., 2006 84.79 90.58 79.19 83.57 73.44 83.17 82.55 88.93 Nivre, 2009 84.2 - - - 75.2 - - - - - - - 83.55 89.30 F.-Gonz. & G.-Rodr., 2012 85.17 90.10 Nivre & McDonald, 2008 86.67 - 81.63 - 75.94 - 84.66 - Martins et al., 2010 - 91.50 - 84.91 - 85.53 - 89.80 14 Wednesday, October 23, 13
  54. 54. SPMRL 2013 Shared Task • Baseline results provided by ClearNLP. Language 5K Full LAS UAS LS LAS UAS LS Arabic 81.72 84.46 93.41 84.19 86.48 94.43 Basque 78.01 84.62 82.71 79.16 85.32 83.63 French 73.39 85.30 81.42 74.51 86.41 82.00 German 82.58 85.36 90.49 86.73 88.80 92.95 Hebrew 75.09 81.74 82.84 - - - Hungarian 81.98 86.09 88.26 82.68 86.56 88.80 Korean 76.28 80.39 87.32 83.55 86.82 92.39 Polish 80.64 88.49 86.47 81.12 89.24 86.59 Swedish 80.96 86.48 85.10 - - - 15 Wednesday, October 23, 13
  55. 55. Conclusion • • Selectional branching - Uses confidence estimates to decide when to employ a beam. Shows comparable accuracy against traditional beam search. Gives faster speed against any other non-greedy parsing. ClearNLP - Provides several NLP tools including morphological analyzer, dependency parser, semantic role labeler, etc. - Webpage: clearnlp.com. 16 Wednesday, October 23, 13

×