Parsing Noun Phrase Structure with C C G
David Vadas and James R. Curran School of Information Technologies University of Sydney NSW 2006, Australia {dvadas1, james}@it.usyd.edu.au

Abstract
Statistical parsing of noun phrase (N P) structure has been hampered by a lack of goldstandard data. This is a significant problem for CCGbank, where binary branching N P derivations are often incorrect, a result of the automatic conversion from the Penn Treebank.

(N (N/N lung) (N (N/N cancer) (N deaths) ) )

This structure is correct for most English N Ps and is the best solution that doesn't require manual reannotation. However, the resulting derivations often contain errors. This can be seen in the previous exWe correct these errors in CCGbank using a gold-standard corpus of N P structure, resultample, where lung cancer should form a coning in a much more accurate corpus. We also stituent, but does not. implement novel N E R features that generalise The first contribution of this paper is to correct the lexical information needed to parse N Ps these CCGbank errors. We apply an automatic conand provide important semantic information. version process using the gold-standard N P data anFinally, evaluating against DepBank demonnotated by Vadas and Curran (2007a). Over a quarstrates the effectiveness of our modified corter of the sentences in CCGbank need to be altered, pus and novel features, with an increase in parser performance of 1.51%. demonstrating the magnitude of the N P problem and how important it is that these errors are fixed. We then run a number of parsing experiments us1 Introduction ing our new version of the CCGbank corpus. In Internal noun phrase (N P) structure is not recovered particular, we implement new features using N E R by a number of widely-used parsers, e.g. Collins tags from the BBN Entity Type Corpus (Weischedel (2003). This is because their training data, the Penn and Brunstein, 2005). These features are targeted at Treebank (Marcus et al., 1993), does not fully anno- improving the recovery of N P structure, increasing tate N P structure. The flat structure described by the parser performance by 0.64% F-score. Penn Treebank can be seen in this example: Finally, we evaluate against DepBank (King et al., (NP (NN lung) (NN cancer) (NNS deaths)) 2003). This corpus annotates internal N P structure, CCGbank (Hockenmaier and Steedman, 2007) is and so is particularly relevant for the changes we the primary English corpus for Combinatory Cate- have made to CCGbank. The C C G parser now recovgorial Grammar (C C G) (Steedman, 2000) and was ers additional structure learnt from our N P corrected created by a semi-automatic conversion from the corpus, increasing performance by 0.92%. Applying Penn Treebank. However, C C G is a binary branch- the N E R features results in a total increase of 1.51%. ing grammar, and as such, cannot leave N P structure This work allows parsers trained on CCGbank to underspecified. Instead, all N Ps were made right- model N P structure accurately, and then pass this branching, as shown in this example: crucial information on to downstream systems. 335
Proceedings of ACL-08: HLT, pages 335­343, Columbus, Ohio, USA, June 2008. c 2008 Association for Computational Linguistics


(a)

(b)

N N /N cotton conj and N /N acetate N N N fibers N /N cotton N /N

N N fibers

N /N [conj ] conj and N /N acetate

Figure 1: (a) Incorrect C C G derivation from Hockenmaier and Steedman (2007) (b) The correct derivation

2 Background
Parsing of N Ps is typically framed as N P bracketing, where the task is limited to discriminating between left and right-branching N Ps of three nouns only: · (crude oil) prices ­ left-branching · world (oil prices) ­ right-branching Lauer (1995) presents two models to solve this problem: the adjacency model, which compares the association strength between words 1­2 to words 2­3; and the dependency model, which compares words 1­2 to words 1­3. Lauer (1995) experiments with a data set of 244 N Ps, and finds that the dependency model is superior, achieving 80.7% accuracy. Most N P bracketing research has used Lauer's data set. Because it is a very small corpus, most approaches have been unsupervised, measuring association strength with counts from a separate large corpus. Nakov and Hearst (2005) use search engine hit counts and extend the query set with typographical markers. This results in 89.3% accuracy. Recently, Vadas and Curran (2007a) annotated internal N P structure for the entire Penn Treebank, providing a large gold-standard corpus for N P bracketing. Vadas and Curran (2007b) carry out supervised experiments using this data set of 36,584 N Ps, outperforming the Collins (2003) parser. The Vadas and Curran (2007a) annotation scheme inserts NML and JJP brackets to describe the correct N P structure, as shown below:
(NP (NML (NN lung) (NN cancer) ) (NNS deaths) )

grammar. Lexical categories (also called supertags) are made up of basic atoms such as S (Sentence) and NP (Noun Phrase), which can be combined to form complex categories. For example, a transitive verb such as bought (as in IBM bought the company) would have the category: (S \NP )/NP . The slashes indicate the directionality of arguments, here two arguments are expected: an N P subject on the left; and an N P object on the right. Once these arguments are filled, a sentence is produced. Categories are combined using combinatory rules such as forward and backward application: X /Y Y X (>) (1) Y X \Y  X (<) (2) Other rules such as composition and type-raising are used to analyse some linguistic constructions, while retaining the canonical categories for each word. This is an advantage of C C G, allowing it to recover long-range dependencies without the need for postprocessing, as is the case for many other parsers. In Section 1, we described the incorrect N P structures in CCGbank, but a further problem that highlights the need to improve N P derivations is shown in Figure 1. When a conjunction occurs in an N P, a non-C C G rule is required in order to reach a parse: conj N  N (3) This rule treats the conjunction in the same manner as a modifier, and results in the incorrect derivation shown in Figure 1(a). Our work creates the correct C C G derivation, shown in Figure 1(b), and removes the need for the grammar rule in (3). Honnibal and Curran (2007) have also made changes to CCGbank, aimed at better differentiating between complements and adjuncts. PropBank (Palmer et al., 2005) is used as a gold-standard to inform these decisions, similar to the way that we use the Vadas and Curran (2007a) data.

We use these brackets to determine new goldstandard C C G derivations in Section 3. 2.1 Combinatory Categorial Grammar

Combinatory Categorial Grammar (C C G) (Steedman, 2000) is a type-driven, lexicalised theory of 336


(a)

(b)

(c)

N N /N lung N /N ca n cer N N deaths ??? lung ???

N ??? ??? ca n cer deaths N /N (N /N )/(N /N ) lung

N N N /N ca n cer deaths

Figure 2: (a) Original right-branching CCGbank (b) Left-branching (c) Left-branching with new supertags

2.2

CCG

parsing

The C & C C C G parser (Clark and Curran, 2007b) is used to perform our experiments, and to evaluate the effect of the changes to CCGbank. The parser uses a two-stage system, first employing a supertagger (Bangalore and Joshi, 1999) to propose lexical categories for each word, and then applying the C K Y chart parsing algorithm. A log-linear model is used to identify the most probable derivation, which makes it possible to add the novel features we describe in Section 4, unlike a P C F G. The C & C parser is evaluated on predicateargument dependencies derived from CCGbank. These dependencies are represented as 5-tuples: hf , f , s, ha , l , where hf is the head of the predicate; f is the supertag of hf ; s describes which argument of f is being filled; ha is the head of the argument; and l encodes whether the dependency is local or long-range. For example, the dependency encoding company as the object of bought (as in IBM bought the company) is represented by: bought, (S \NP1 )/NP2 , 2, company , - (4)

tokens. For example, we would insert the NML bracket shown below:
(NP (DT a) (-LRB- -LRB-) (NML (RB very) (JJ negative) ) (-RRB- -RRB-) (NN reaction) )

This simple heuristic captures N P structure not explicitly annotated by Vadas and Curran (2007a). The conversion algorithm applies the following steps for each NML or JJP bracket: 1. Identify the CCGbank lowest spanning node, the lowest constituent that covers all of the words in the NML or JJP bracket; 2. flatten the lowest spanning node, to remove the right-branching structure; 3. insert new left-branching structure; 4. identify heads; 5. assign supertags; 6. generate new dependencies. As an example, we will follow the conversion process for the NML bracket below:
(NP (NML (NN lung) (NN cancer) ) (NNS deaths) )

This is a local dependency, where company is filling the second argument slot, the object.

3 Conversion Process
This section describes the process of converting the Vadas and Curran (2007a) data to C C G derivations. The tokens dominated by NML and JJP brackets in the source data are formed into constituents in the corresponding CCGbank sentence. We generate the two forms of output that CCGbank contains: AUTO files, which represent the tree structure of each sentence; and PARG files, which list the word­word dependencies (Hockenmaier and Steedman, 2005). We apply one preprocessing step on the Penn Treebank data, where if multiple tokens are enclosed by brackets, then a NML node is placed around those 337

The corresponding lowest spanning node, which incorrectly has cancer deaths as a constituent, is shown in Figure 2(a). To flatten the node, we recursively remove brackets that partially overlap the NML bracket. Nodes that don't overlap at all are left intact. This process results in a list of nodes (which may or may not be leaves), which in our example is [lung, cancer, deaths]. We then insert the correct left-branching structure, shown in Figure 2(b). At this stage, the supertags are still incomplete. Heads are then assigned using heuristics adapted from Hockenmaier and Steedman (2007). Since we are applying these to CCGbank N P structures rather than the Penn Treebank, the P O S tag based heuristics are sufficient to determine heads accurately.


Finally, we assign supertags to the new structure. We want to make the minimal number of changes to the entire sentence derivation, and so the supertag of the dominating node is fixed. Categories are then propagated recursively down the tree. For a node with category X , its head child is also given the category X . The non-head child is always treated as an adjunct, and given the category X /X or X \X as appropriate. Figure 2(c) shows the final result of this step for our example. 3.1 Dependency generation

during the head-finding stage, and then assigned the supertag dominating the entire coordination. Intervening non-conjunct nodes are given the same category with the conj feature, resulting in a derivation that can be parsed with the standard CCGbank binary coordination rules: conj X  X [conj ] X X [conj ]  X (8) (9)

The changes described so far have generated the new tree structure, but the last step is to generate new dependencies. We recursively traverse the tree, at each level creating a dependency between the heads of the left and right children. These dependencies are never long-range, and therefore easy to deal with. We may also need to change dependencies reaching from inside to outside the N P, if the head(s) of the N P have changed. In these cases we simply replace the old head(s) with the new one(s) in the relevant dependencies. The number of heads may change because we now analyse conjunctions correctly. In our example, the original dependencies were: lung, N /N1 , 1, deaths, - cancer, N /N1 , 1, deaths, - while after the conversion process, (5) becomes: lung, (N /N1 )/(N /N )2 , 2, cancer, - (7) (5) (6)

The derivation in Figure 1(b) is produced by these corrections to coordination derivations. As a result, applications of the non-C C G rule shown in (3) have been reduced from 1378 to 145 cases. Some P O S tags require special behaviour. Determiners and possessive pronouns are both usually given the supertag NP [nb ]/N , and this should not be changed by the conversion process. Accordingly, we do not alter tokens with P O S tags of DT and PRP$. Instead, their sibling node is given the category N and their parent node is made the head. The parent's sibling is then assigned the appropriate adjunct category (usually NP \NP ). Tokens with punctuation P O S tags1 do not have their supertag changed either. Finally, there are cases where the lowest spanning node covers a constituent that should not be changed. For example, in the following N P:
(NP (NML (NN lower) (NN court) ) (JJ final) (NN ruling) )

with the original CCGbank lowest spanning node:
(N (N/N lower) (N (N/N court) (N (N/N final) (N ruling) ) ) )

To determine that the conversion process worked correctly, we manually inspected its output for unique tree structures in Sections 00­07. This identified problem cases to correct, such as those described in the following section. 3.2 Exceptional cases

the final ruling node should not be altered. It may seem trivial to process in this case, but consider a similarly structured N P: lower court
ruling that the U.S. can bar the use of... Our minimalist approach avoids reanalysing

Firstly, when the lowest spanning node covers the NML or JJP bracket exactly, no changes need to be made to CCGbank. These cases occur when CCGbank already received the correct structure during the original conversion process. For example, brackets separating a possessive from its possessor were detected automatically. A more complex case is conjunctions, which do not follow the simple head/adjunct method of assigning supertags. Instead, conjuncts are identified 338

the many linguistic constructions that can be dominated by N Ps, as this would reinvent the creation of CCGbank. As a result, we only flatten those constituents that partially overlap the NML or JJP bracket. The existing structure and dependencies of other constituents are retained. Note that we are still converting every NML and JJP bracket, as even in the subordinate clause example, only the structure around lower court needs to be altered.
1

period, comma, colon, and left and right bracket.


the

world 's N NP NP NP
> >

largest

aid

donor

the NP [nb ]/N NP

world N
>

's

largest

aid

donor N N
> >

NP [nb ]/N N /N N NP \NP NP \NP NP \NP

(NP [nb ]/N )\NP N /N N /N
<

NP [nb ]/N
< <

N NP
>

NP ( a)

<

(b ) Figure 3: CCGbank derivations for possessives

Possessive Left child contains DT/PRP$ Couldn't assign to non-leaf Conjunction Automatic conversion was correct Entity with internal brackets DT NML/JJP bracket is an error Other Total
Table 1: Manual analysis

# 224 87 66 35 26 23 22 12 17 512

% 43.75 16.99 12.89 6.84 5.08 4.49 4.30 2.34 3.32 100.00

3.3

Manual annotation

A handful of problems that occurred during the conversion process were corrected manually. The first indicator of a problem was the presence of a possessive. This is unexpected, because possessives were already bracketed properly when CCGbank was originally created (Hockenmaier, 2003, §3.6.4). Secondly, a non-flattened node should not be assigned a supertag that it did not already have. This is because, as described previously, a non-leaf node could dominate any kind of structure. Finally, we expect the lowest spanning node to cover only the NML or JJP bracket and one more constituent to the right. If it doesn't, because of unusual punctuation or an incorrect bracket, then it may be an error. In all these cases, which occur throughout the corpus, we manually analysed the derivation and fixed any errors that were observed. 512 cases were flagged by this approach, or 1.90% of the 26,993 brackets converted to C C G. Table 1 shows the causes of these problems. The most common cause of errors was possessives, as the con339

version process highlighted a number of instances where the original CCGbank analysis was incorrect. An example of this error can be seen in Figure 3(a), where the possessive doesn't take any arguments. Instead, largest aid donor incorrectly modifies the N P one word at a time. The correct derivation after manual analysis is in (b). The second-most common cause occurs when there is apposition inside the N P. This can be seen in Figure 4. As there is no punctuation on which to coordinate (which is how CCGbank treats most appositions) the best derivation we can obtain is to have Victor Borge modify the preceding N P. The final step in the conversion process was to validate the corpus against the C C G grammar, first by those productions used in the existing CCGbank, and then against those actually licensed by C C G (with pre-existing ungrammaticalities removed). Sixteen errors were identified by this process and subsequently corrected by manual analysis. In total, we have altered 12,475 CCGbank sentences (25.5%) and 20,409 dependencies (1.95%).

4

NER

features

Named entity recognition (N E R) provides information that is particularly relevant for N P parsing, simply because entities are nouns. For example, knowing that Air Force is an entity tells us that Air Force contract is a left-branching N P. Vadas and Curran (2007a) describe using N E tags during the annotation process, suggesting that N E Rbased features will be helpful in a statistical model. There has also been recent work combining N E R and parsing in the biomedical field. Lewin (2007) experiments with detecting base-N Ps using N E R information, while Buyko et al. (2007) use a C R F to identify


a

guest comedian Victor Borge N /N N /N N N N NP ( a) N
> >

a

guest comedian N
>

Victor NP \ NP

Borge
>

NP [nb ]/N N /N

NP [nb ]/N N /N N NP
> >

(NP \NP )/(NP \NP ) NP \NP
>

NP (b )

<

Figure 4: CCGbank derivations for apposition with DT

coordinate structure in biological named entities. We draw N E tags from the BBN Entity Type Corpus (Weischedel and Brunstein, 2005), which describes 28 different entity types. These include the standard person, location and organization classes, as well person descriptions (generally occupations), NORP (National, Other, Religious or Political groups), and works of art. Some classes also have finer-grained subtypes, although we use only the coarse tags in our experiments. Clark and Curran (2007b) has a full description of the C & C parser's pre-existing features, to which we have added a number of novel N E R-based features. Many of these features generalise the head words and/or P O S tags that are already part of the feature set. The results of applying these features are described in Sections 5.3 and 6. The first feature is a simple lexical feature, describing the N E tag of each token in the sentence. This feature, and all others that we describe here, are not active when the N E tag(s) are O, as there is no N E R information from tokens that are not entities. The next group of features is based on the local tree (a parent and two child nodes) formed by every grammar rule application. We add a feature where the rule being applied is combined with the parent's N E tag. For example, when joining two constituents2 : five, CD, CARD, N /N and Europeans, NNPS, NORP, N , the feature is: N  N /N N + NORP as the head of the constituent is Europeans. In the same way, we implement features that combine the grammar rule with the child nodes. There are already features in the model describing each combination of the children's head words and P O S tags, which we extend to include combinations with
2

the N E tags. Using the same example as above, one of the new features would be: N  N /N N + CARD + NORP The last group of features is based on the N E category spanned by each constituent. We identify constituents that dominate tokens that all have the same N E tag, as these nodes will not cause a "crossing bracket" with the named entity. For example, the constituent Force contract, in the N P Air Force contract, spans two different N E tags, and should be penalised by the model. Air Force, on the other hand, only spans ORG tags, and should be preferred accordingly. We also take into account whether the constituent spans the entire named entity. Combining these nodes with others of different N E tags should not be penalised by the model, as the N E must combine with the rest of the sentence at some point. These N E spanning features are implemented as the grammar rule in combination with the parent node or the child nodes. For the former, one feature is active when the node spans the entire entity, and another is active in other cases. Similarly, there are four features for the child nodes, depending on whether neither, the left, the right or both nodes span the entire N E. As an example, if the Air Force constituent were being joined with contract, then the child feature would be: N  N /N N + LEFT + ORG + O assuming that there are more O tags to the right.

5 Experiments
Our experiments are run with the C & C C C G parser (Clark and Curran, 2007b), and will evaluate the changes made to CCGbank, as well as the effectiveness of the N E R features. We train on Sections 0221, and test on Section 00.

These 4-tuples are the node's head, P O S, N E , and supertag.

340


PREC

RECALL

F-SCORE

PREC

RECALL

F-SCORE

Original N P corrected

91.85 91.22

92.67 92.08

92.26 91.65

Original N P corrected

83.65 83.31

82.81 82.33

83.23 82.82

Table 2: Supertagging results
PREC RECALL F-SCORE

Table 4: Parsing results with automatic P O S tags
F-SCORE

Original N P corrected

85.34 85.08

84.55 84.17

84.94 84.63

PREC

RECALL

Original N P corrected

86.00 85.71

85.15 84.83

85.58 85.27

Table 3: Parsing results with gold-standard P O S tags

Table 5: Parsing results with N E R features

5.1

Supertagging way, as N P dependencies remain undifferentiated in parser output. The result is a recall of 77.03%, which is noticeably lower than the overall figure. We have also experimented with using automatically assigned P O S tags. These tags are accurate with an F-score of 96.34%, with precision 96.20% and recall 96.49%. Table 4 shows that, unsurprisingly, performance is lower without the goldstandard data. The N P corrected model drops an additional 0.1% F-score over the original model, suggesting that P O S tags are particularly important for recovering internal N P structure. Evaluating N P dependencies only, in the same manner as before, results in a recall figure of 75.21%.

Before we begin full parsing experiments, we evaluate on the supertagger alone. The supertagger is an important stage of the C C G parsing process, its results will affect performance in later experiments. Table 2 shows that F-score has dropped by 0.61%. This is not surprising, as the conversion process has increased the ambiguity of supertags in N Ps. Previously, a bare N P could only have a sequence of N /N tags followed by a final N . There are now more complex possibilities, equal to the Catalan number of the length of the N P. 5.2 Initial parsing results

We now compare parser performance on our N P corrected version of the corpus to that on original CCGbank. We are using the normal-form parser model and report labelled precision, recall and F-score for all dependencies. The results are shown in Table 3. The F-score drops by 0.31% in our new version of the corpus. However, this comparison is not entirely fair, as the original CCGbank test data does not include the N P structure that the N P corrected model is being evaluated on. Vadas and Curran (2007a) experienced a similar drop in performance on Penn Treebank data, and noted that the F-score for NML and JJP brackets was about 20% lower than the overall figure. We suspect that a similar effect is causing the drop in performance here. Unfortunately, there are no explicit NML and JJP brackets to evaluate on in the C C G corpus, and so an N P structure only figure is difficult to compute. Recall can be calculated by marking those dependencies altered in the conversion process, and evaluating only on them. Precision cannot be measured in this 341

5.3

NER

features results

Table 5 shows the results of adding the N E R features we described in Section 4. Performance has increased by 0.64% on both versions of the corpora. It is surprising that the N P corrected increase is not larger, as we would expect the features to be less effective on the original CCGbank. This is because incorrect right-branching N Ps such as Air Force contract would introduce noise to the N E R features. Table 6 presents the results of using automatically assigned P O S and N E tags, i.e. parsing raw text. The N E R tagger achieves 84.45% F-score on all non-O classes, with precision being 78.35% and recall 91.57%. We can see that parsing F-score has dropped by about 2% compared to using goldstandard P O S and N E R data, however, the N E R features still improve performance by about 0.3%.


PREC

RECALL

F-SCORE

PREC

RECALL

F-SCORE

Original N P corrected

83.92 83.62

83.06 82.65

83.49 83.14

Original N P corrected

86.86 87.97

81.61 82.54

84.15 85.17

Table 6: Parsing results with automatic P O S and N E tags

Table 7: DepBank gold-standard evaluation
PREC RECALL F-SCORE

6 DepBank evaluation
One problem with the evaluation in the previous section, is that the original CCGbank is not expected to recover internal N P structure, making its task easier and inflating its performance. To remove this variable, we carry out a second evaluation against the Briscoe and Carroll (2006) reannotation of DepBank (King et al., 2003), as described in Clark and Curran (2007a). Parser output is made similar to the grammatical relations (G Rs) of the Briscoe and Carroll (2006) data, however, the conversion remains complex. Clark and Curran (2007a) report an upper bound on performance, using gold-standard CCGbank dependencies, of 84.76% F-score. This evaluation is particularly relevant for N Ps, as the Briscoe and Carroll (2006) corpus has been annotated for internal N P structure. With our new version of CCGbank, the parser will be able to recover these G Rs correctly, where before this was unlikely. Firstly, we show the figures achieved using goldstandard CCGbank derivations in Table 7. In the N P corrected version of the corpus, performance has increased by 1.02% F-score. This is a reversal of the results in Section 5, and demonstrates that correct N P structure improves parsing performance, rather than reduces it. Because of this increase to the upper bound of performance, we are now even closer to a true formalism-independent evaluation. We now move to evaluating the C & C parser itself and the improvement gained by the N E R features. Table 8 show our results, with the N P corrected version outperforming original CCGbank by 0.92%. Using the N E R features has also caused an increase in F-score, giving a total improvement of 1.51%. These results demonstrate how successful the correcting of N Ps in CCGbank has been. Furthermore, the performance increase of 0.59% on the N P corrected corpus is more than the 0.25% increase on the original. This demonstrates that N E R features are particularly helpful for N P structure. 342

Original N P corrected Original, N E R N P corrected, N E R

82.57 83.53 82.87 84.12

81.29 82.15 81.49 82.75

81.92 82.84 82.17 83.43

Table 8: DepBank evaluation results

7 Conclusion
The first contribution of this paper is the application of the Vadas and Curran (2007a) data to Combinatory Categorial Grammar. Our experimental results have shown that this more accurate representation of CCGbank's N P structure increases parser performance. Our second major contribution is the introduction of novel N E R features, a source of semantic information previously unused in parsing. As a result of this work, internal N P structure is now recoverable by the C & C parser, a result demonstrated by our total performance increase of 1.51% F-score. Even when parsing raw text, without gold standard P O S and N E R tags, our approach has resulted in performance gains. In addition, we have made possible further increases to N P structure accuracy. New features can now be implemented and evaluated in a C C G parsing context. For example, bigram counts from a very large corpus have already been used in N P bracketing, and could easily be applied to parsing. Similarly, additional supertagging features can now be created to deal with the increased ambiguity in N Ps. Downstream N L P components can now exploit the crucial information in N P structure.

Acknowledgements
We would like to thank Mark Steedman and Matthew Honnibal for help with converting the N P data to C C G; and the anonymous reviewers for their helpful feedback. This work has been supported by the Australian Research Council under Discovery Project DP0665973.


References
Srinivas Bangalore and Aravind Joshi. 1999. Supertagging: An approach to almost parsing. Computational Linguistics, 25(2):237­265. Ted Briscoe and John Carroll. 2006. Evaluating the accuracy of an unlexicalized statistical parser on the PARC DepBank. In Proceedings of the COLING/ACL 2006 Main Conference Poster Sessions, pages 41­48. Sydney, Australia. Ekaterina Buyko, Katrin Tomanek, and Udo Hahn. 2007. Resolution of coordination ellipses in biological named entities with conditional random fields. In Proceedings of the 10th Conference of the Pacific Association for Computational Linguistics (PACLING-2007), pages 163­171. Melbourne, Australia. Stephen Clark and James R. Curran. 2007a. Formalismindependent parser evaluation with CCG and DepBank. In Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics (ACL07), pages 248­255. Prague, Czech Republic. Stephen Clark and James R. Curran. 2007b. Widecoverage efficient statistical parsing with CCG and log-linear models. Computational Linguistics, 33(4):493­552. Michael Collins. 2003. Head-driven statistical models for natural language parsing. Computational Linguistics, 29(4):589­637. Julia Hockenmaier. 2003. Data and Models for Statistical Parsing with Combinatory Categorial Grammar. Ph.D. thesis, University of Edinburgh. Julia Hockenmaier and Mark Steedman. 2005. CCGbank manual. Technical Report MS-CIS-05-09, Department of Computer and Information Science, University of Pennsylvania. Julia Hockenmaier and Mark Steedman. 2007. CCGbank: a corpus of CCG derivations and dependency structures extracted from the Penn Treebank. Computational Linguistics, 33(3):355­396. Matthew Honnibal and James R. Curran. 2007. Improving the complement/adjunct distinction in CCGbank. In Proceedings of the 10th Conference of the Pacific Association for Computational Linguistics (PACLING-07), pages 210­217. Melbourne, Australia. Tracy Holloway King, Richard Crouch, Stefan Riezler, Mary Dalrymple, and Ronald M. Kaplan. 2003. The PARC700 dependency bank. In Proceedings of the 4th International Workshop on Linguistically Interpreted Corpora (LINC-03). Budapest, Hungary. Mark Lauer. 1995. Corpus statistics meet the compound noun: Some empirical results. In Proceedings of the

33rd Annual Meeting of the Association for Computational Linguistics, pages 47­54. Cambridge, MA. Ian Lewin. 2007. BaseNPs that contain gene names: domain specificity and genericity. In Biological, translational, and clinical language processing workshop, pages 163­170. Prague, Czech Republic. Mitchell Marcus, Beatrice Santorini, and Mary Marcinkiewicz. 1993. Building a large annotated corpus of English: The Penn Treebank. Computational Linguistics, 19(2):313­330. Preslav Nakov and Marti Hearst. 2005. Search engine statistics beyond the n-gram: Application to noun compound bracketing. In Proceedings of the 9th Conference on Computational Natural Language Learning (CoNLL-05), pages 17­24. Ann Arbor, MI. Martha Palmer, Daniel Gildea, and Paul Kingsbury. 2005. The proposition bank: An annotated corpus of semantic roles. Computational Linguistics, 31(1):71­106. Mark Steedman. 2000. The Syntactic Process. MIT Press, Cambridge, MA. David Vadas and James R. Curran. 2007a. Adding noun phrase structure to the Penn Treebank. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics (ACL-07), pages 240­247. Prague, Czech Republic. David Vadas and James R. Curran. 2007b. Large-scale supervised models for noun phrase bracketing. In Proceedings of the 10th Conference of the Pacific Association for Computational Linguistics (PACLING-2007), pages 104­112. Melbourne, Australia. Ralph Weischedel and Ada Brunstein. 2005. BBN pronoun coreference and entity type corpus. Technical report.

343