Linguistik-Klassifikation
Refine
Year of publication
Document Type
- Preprint (46)
- Conference Proceeding (17)
- Book (5)
- Part of a Book (5)
- Article (4)
- Working Paper (3)
Language
- English (80) (remove)
Has Fulltext
- yes (80) (remove)
Is part of the Bibliography
- no (80)
Keywords
- Computerlinguistik (28)
- Japanisch (15)
- Deutsch (13)
- Maschinelle Übersetzung (8)
- Multicomponent Tree Adjoining Grammar (8)
- Syntaktische Analyse (8)
- Satzanalyse (4)
- Semantik (4)
- German (3)
- Grammatik (3)
Institute
- Extern (80) (remove)
We present an architecture for the integration of shallow and deep NLP components which is aimed at flexible combination of different language technologies for a range of practical current and future applications. In particular, we describe the integration of a high-level HPSG parsing system with different high-performance shallow components, ranging from named entity recognition to chunk parsing and shallow clause recognition. The NLP components enrich a representation of natural language text with layers of new XML meta-information using a single shared data structure, called the text chart. We describe details of the integration methods, and show how information extraction and language checking applications for realworld German text benefit from a deep grammatical analysis.
Hybrid robust deep and shallow semantic processing for creativity support in document production
(2004)
The research performed in the DeepThought project (http://www.project-deepthought.net) aims at demonstrating the potential of deep linguistic processing if added to existing shallow methods that ensure robustness. Classical information retrieval is extended by high precision concept indexing and relation detection. We use this approach to demonstrate the feasibility of three ambitious applications, one of which is a tool for creativity support in document production and collective brainstorming. This application is described in detail in this paper. Common to all three applications, and the basis for their development is a platform for integrated linguistic processing. This platform is based on a generic software architecture that combines multiple NLP components and on robust minimal recursive semantics (RMRS) as a uniform representation language.
In this paper we show an approach to the customization of GermaNet to the German HPSG grammar lexicon developed in the Verbmobil project. GermaNet has a broad coverage of the German base vocabulary and fine-grained semantic classification; while the HPSG grammar lexicon is comparatively small und has a coarse-grained semantic classification. In our approach, we have developed a mapping algorithm to relate the synsets in GermaNet with the semantic sorts in HPSG. The evaluation result shows that this approach is useful for the lexical extension of our deep grammar development to cope with real-world text understanding.
Particles fullfill several distinct central roles in the Japanese language. They can mark arguments as well as adjuncts, can be functional or have semantic functions. There is, however, no straightforward matching from particles to functions, as, e.g., 'ga' can mark the subject, the object or the adjunct of a sentence. Particles can cooccur. Verbal arguments that could be identified by particles can be eliminated in the Japanese sentence. And finally, in spoken language particles are often omitted. A proper treatment of particles is thus necessary to make an analysis of Japanese sentences possible. Our treatment is based on an empirical investigation of 800 dialogues. We set up a type hierarchy of particles motivated by their subcategorizational and modificational behaviour. This type hierarchy is part of the Japanese syntax in VERBMOBIL.
Preferences and defaults for definiteness and number in japanese to german machine translation
(1996)
A significant problem when translating Japanese dialogues into German is the missing information on number and definiteness in the Japanese analysis output. The integration of the search for such information into the transfer process provides an efficient solution. General transfer includes conditions to make it possible to consider external knowledge. Thereby, grammatical and lexical knowledge of the source language, knowledge of lexical restrictions on the target language, domain knowledge and discourse knowledge are accessible.
A comprehensive investigation of Japanese particle was missing up to now. General implications were set up without the fact that a comprehensive analysis was carried out. [...] We offer a lexicalist treatment of the problem. Instead of assuming different phrase structure rules we state a type hierarchy of Japanese particles. This makes a uniform treatment of phrase structure as well as a differentiation of subcategorization patterns possible.
We present a solution for the representation of Japanese honorifical information in the HPSG framework. Basically, there are three dimensions of honorification. We show that a treatment is necessary that involves both the syntactic and the contextual level of information. The japanese grammar is part of a machine translation system.
The research performed in the DeepThought project aims at demonstrating the potential of deep linguistic processing if combined with shallow methods for robustness. Classical information retrieval is extended by high precision concept indexing and relation detection. On the basis of this approach, the feasibility of three ambitious applications will be demonstrated, namely: precise information extraction for business intelligence; email response management for customer relationship management; creativity support for document production and collective brainstorming. Common to these applications, and the basis for their development is the XML-based, RMRS-enabled core architecture framework that will be described in detail in this paper. The framework is not limited to the applications envisaged in the DeepThought project, but can also be employed e.g. to generate and make use of XML standoff annotation of documents and linguistic corpora, and in general for a wide range of NLP-based applications and research purposes.
In this paper we describe SOBA, a sub-component of the SmartWeb multi-modal dialog system. SOBA is a component for ontologybased information extraction from soccer web pages for automatic population of a knowledge base that can be used for domainspecific question answering. SOBA realizes a tight connection between the ontology, knowledge base and the information extraction component. The originality of SOBA is in the fact that it extracts information from heterogeneous sources such as tabular structures, text and image captions in a semantically integrated way. In particular, it stores extracted information in a knowledge base, and in turn uses the knowledge base to interpret and link newly extracted information with respect to already existing entities.
This demo abstract describes the SmartWeb Ontology-based Information Extraction System (SOBIE). A key feature of SOBIE is that all information is extracted and stored with respect to the SmartWeb ontology. In this way, other components of the systems, which use the same ontology, can access this information in a straightforward way. We will show how information extracted by SOBIE is visualized within its original context, thus enhancing the browsing experience of the end user.
Dialogue acts in Verbmobil 2
(1998)
This report describes the dialogue phases and the second edition dialogue acts which are used in the VERBMOBIL 2 project [...]. While in the first project phase the scenario was restricted to appointment scheduling dialogues, it has been extended to travel planning in the second phase with appointment scheduling being only a part of the new scenario.
The Deep Linguistic Processing with HPSG Initiative (DELH-IN) provides the infrastructure needed to produce open-source semantic transfer-based machine translation systems. We have made available a prototype Japanese-English machine translation system built from existing resources include parsers, generators, bidirectional grammars and a transfer engine.
While the sortal constraints associated with Japanese numeral classifiers are well-studied, less attention has been paid to the details of their syntax. We describe an analysis implemented within a broad-coverage HPSG that handles an intricate set of numeral classifier construction types and compositionally relates each to an appropriate semantic representation, using Minimal Recursion Semantics.
While the sortal constraints associated with Japanese numeral classifiers are wellstudied, less attention has been paid to the details of their syntax. We describe an analysis implemented within a broadcoverage HPSG that handles an intricate set of numeral classifier construction types and compositionally relates each to an appropriate semantic representation, using Minimal Recursion Semantics.
We present an effort for the development of multilingual named entity grammars in a unification-based finite-state formalism (SProUT). Following an extended version of the MUC7 standard, we have developed Named Entity Recognition grammars for German, Chinese, Japanese, French, Spanish, English, and Czech. The grammars recognize person names, organizations, geographical locations, currency, time and date expressions. Subgrammars and gazetteers are shared as much as possible for the grammars of the different languages. Multilingual corpora from the business domain are used for grammar development and evaluation. The annotation format (named entity and other linguistic information) is described. We present an evaluation tool which provides detailed statistics and diagnostics, allows for partial matching of annotations, and supports user-defined mappings between different annotation and grammar output formats.
MED (Media EDitor) is a program designed to facilitate the transcription of digitized soundfiles into textfiles. It was written by Hans Drexler and Daan Broeder, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands. [...] The aim of MED is to facilitate the transcription of sound into text using a single program. It works on the principle of the coexistence and interaction of two basic elements, the waveform display window and the text window. [...] This means that you no longer need to use both a sound editor and a word processor at the same time in order to transcribe digitized speech files. Instead, you can directly type the sound you hear (and see) via MED into the text window. Furthermore, you can directly link sound portions of the waveform display window to text portions of the text window, so that you can easily locate and listen to the original source of your transcription once the links have been set. In this function the waveform display window and the text window virtually interact with each other.
This paper investigates the class of Tree-Tuple MCTAG with Shared Nodes, TT-MCTAG for short, an extension of Tree Adjoining Grammars that has been proposed for natural language processing, in particular for dealing with discontinuities and word order variation in languages such as German. It has been shown that the universal recognition problem for this formalism is NP-hard, but so far it was not known whether the class of languages generated by TT-MCTAG is included in PTIME. We provide a positive answer to this question, using a new characterization of TT-MCTAG.
Multicomponent Tree Adjoining Grammars (MCTAGs) are a formalism that has been shown to be useful for many natural language applications. The definition of non-local MCTAG however is problematic since it refers to the process of the derivation itself: a simultaneity constraint must be respected concerning the way the members of the elementary tree sets are added. Looking only at the result of a derivation (i.e., the derived tree and the derivation tree), this simultaneity is no longer visible and therefore cannot be checked. I.e., this way of characterizing MCTAG does not allow to abstract away from the concrete order of derivation. In this paper, we propose an alternative definition of MCTAG that characterizes the trees in the tree language of an MCTAG via the properties of the derivation trees (in the underlying TAG) the MCTAG licences. We provide similar characterizations for various types of MCTAG. These characterizations give a better understanding of the formalisms, they allow a more systematic comparison of different types of MCTAG, and, furthermore, they can be exploited for parsing.
In this paper, we argue that difficulties in the definition of coreference itself contribute to lower inter-annotator agreement in certain cases. Data from a large referentially annotated corpus serves to corroborate this point, using a quantitative investigation to assess which effects or problems are likely to be the most prominent. Several examples where such problems occur are discussed in more detail, and we then propose a generalisation of Poesio, Reyle and Stevenson’s Justified Sloppiness Hypothesis to provide a unified model for these cases of disagreement and argue that a deeper understanding of the phenomena involved allows to tackle problematic cases in a more principled fashion than would be possible using only pre-theoretic intuitions.
When a statistical parser is trained on one treebank, one usually tests it on another portion of the same treebank, partly due to the fact that a comparable annotation format is needed for testing. But the user of a parser may not be interested in parsing sentences from the same newspaper all over, or even wants syntactic annotations for a slightly different text type. Gildea (2001) for instance found that a parser trained on the WSJ portion of the Penn Treebank performs less well on the Brown corpus (the subset that is available in the PTB bracketing format) than a parser that has been trained only on the Brown corpus, although the latter one has only half as many sentences as the former. Additionally, a parser trained on both the WSJ and Brown corpora performs less well on the Brown corpus than on the WSJ one. This leads us to the following questions that we would like to address in this paper: - Is there a difference in usefulness of techniques that are used to improve parser performance between the same-corpus and the different-corpus case? - Are different types of parsers (rule-based and statistical) equally sensitive to corpus variation? To achieve this, we compared the quality of the parses of a hand-crafted constraint-based parser and a statistical PCFG-based parser that was trained on a treebank of German newspaper text.
In the past, a divide could be seen between ’deep’ parsers on the one hand, which construct a semantic representation out of their input, but usually have significant coverage problems, and more robust parsers on the other hand, which are usually based on a (statistical) model derived from a treebank and have larger coverage, but leave the problem of semantic interpretation to the user. More recently, approaches have emerged that combine the robustness of datadriven (statistical) models with more detailed linguistic interpretation such that the output could be used for deeper semantic analysis. Cahill et al. (2002) use a PCFG-based parsing model in combination with a set of principles and heuristics to derive functional (f-)structures of Lexical-Functional Grammar (LFG). They show that the derived functional structures have a better quality than those generated by a parser based on a state-of-the-art hand-crafted LFG grammar. Advocates of Dependency Grammar usually point out that dependencies already are a semantically meaningful representation (cf. Menzel, 2003). However, parsers based on dependency grammar normally create underspecified representations with respect to certain phenomena such as coordination, apposition and control structures. In these areas they are too "shallow" to be directly used for semantic interpretation. In this paper, we adopt a similar approach to Cahill et al. (2002) using a dependency-based analysis to derive functional structure, and demonstrate the feasibility of this approach using German data. A major focus of our discussion is on the treatment of coordination and other potentially underspecified structures of the dependency data input. F-structure is one of the two core levels of syntactic representation in LFG (Bresnan, 2001). Independently of surface order, it encodes abstract syntactic functions that constitute predicate argument structure and other dependency relations such as subject, predicate, adjunct, but also further semantic information such as the semantic type of an adjunct (e.g. directional). Normally f-structure is captured as a recursive attribute value matrix, which is isomorphic to a directed graph representation. Figure 5 depicts an example target f-structure. As mentioned earlier, these deeper-level dependency relations can be used to construct logical forms as in the approaches of van Genabith and Crouch (1996), who construct underspecified discourse representations (UDRSs), and Spreyer and Frank (2005), who have robust minimal recursion semantics (RMRS) as their target representation. We therefore think that f-structures are a suitable target representation for automatic syntactic analysis in a larger pipeline of mapping text to interpretation. In this paper, we report on the conversion from dependency structures to fstructure. Firstly, we evaluate the f-structure conversion in isolation, starting from hand-corrected dependencies based on the TüBa-D/Z treebank and Versley (2005)´s conversion. Secondly, we start from tokenized text to evaluate the combined process of automatic parsing (using Foth and Menzel (2006)´s parser) and f-structure conversion. As a test set, we randomly selected 100 sentences from TüBa-D/Z which we annotated using a scheme very close to that of the TiGer Dependency Bank (Forst et al., 2004). In the next section, we sketch dependency analysis, the underlying theory of our input representations, and introduce four different representations of coordination. We also describe Weighted Constraint Dependency Grammar (WCDG), the dependency parsing formalism that we use in our experiments. Section 3 characterises the conversion of dependencies to f-structures. Our evaluation is presented in section 4, and finally, section 5 summarises our results and gives an overview of problems remaining to be solved.
This paper is concerned with the tagging of spatial expressions in German newspaper articles, assigning a meaning to the expression and classifying the usages of the spatial expression and linking the derived referent to an event description. In our system, we implemented the activation of concepts in a very simple fashion, a concept is activated once (with a cost depending on the item that activated it) and is left activated thereafter. As an example, a city also activates the nodes for the region and the country it is part of, so that cities from one country are chosen over cities from different countries. A test corpus of 12 German newspaper articles was tested regarding several disambiguation strategies. Disambiguation was carried out via a beam search to find an approximately cost-optimal solution for the conflict set of potential grounding candidates for the tagged spatial expression. Test showed that the disambiguation strategies improved accuracy significantly.
Using a qualitative analysis of disagreements from a referentially annotated newspaper corpus, we show that, in coreference annotation, vague referents are prone to greater disagreement. We show how potentially problematic cases can be dealt with in a way that is practical even for larger-scale annotation, considering a real-world example from newspaper text.
We investigate methods to improve the recall in coreference resolution by also trying to resolve those definite descriptions where no earlier mention of the referent shares the same lexical head (coreferent bridging). The problem, which is notably harder than identifying coreference relations among mentions which have the same lexical head, has been tackled with several rather different approaches, and we attempt to provide a meaningful classification along with a quantitative comparison. Based on the different merits of the methods, we discuss possibilities to improve them and show how they can be effectively combined.
In recent years, research in parsing has extended in several new directions. One of these directions is concerned with parsing languages other than English. Treebanks have become available for many European languages, but also for Arabic, Chinese, or Japanese. However, it was shown that parsing results on these treebanks depend on the types of treebank annotations used. Another direction in parsing research is the development of dependency parsers. Dependency parsing profits from the non-hierarchical nature of dependency relations, thus lexical information can be included in the parsing process in a much more natural way. Especially machine learning based approaches are very successful (cf. e.g.). The results achieved by these dependency parsers are very competitive although comparisons are difficult because of the differences in annotation. For English, the Penn Treebank has been converted to dependencies. For this version, Nivre et al. report an accuracy rate of 86.3%, as compared to an F-score of 92.1 for Charniaks parser. The Penn Chinese Treebank is also available in a constituent and a dependency representations. The best results reported for parsing experiments with this treebank give an F-score of 81.8 for the constituent version and 79.8% accuracy for the dependency version. The general trend in comparisons between constituent and dependency parsers is that the dependency parser performs slightly worse than the constituent parser. The only exception occurs for German, where F-scores for constituent plus grammatical function parses range between 51.4 and 75.3, depending on the treebank, NEGRA or TüBa-D/Z. The dependency parser based on a converted version of Tüba-D/Z, in contrast, reached an accuracy of 83.4%, i.e. 12 percent points better than the best constituent analysis including grammatical functions.
This paper profiles significant differences in syntactic distribution and differences in word class frequencies for two treebanks of spoken and written German: the TüBa-D/S, a treebank of transliterated spontaneous dialogues, and the TüBa-D/Z treebank of newspaper articles published in the German daily newspaper die tageszeitung´(taz). The approach can be used more generally as a means of distinguishing and classifying language corpora of different genres.
This paper profiles significant differences in syntactic distribution and differences in word class frequencies for two treebanks of spoken and written German: the TüBa-D/S, a treebank of transliterated spontaneous dialogs, and the TüBa-D/Z treebank of newspaper articles published in the German daily newspaper ´die tageszeitung´(taz). The approach can be used more generally as a means of distinguishing and classifying language corpora of different genres.
Quantitative evaluation of parsers has traditionally centered around the PARSEVAL measures of crossing brackets, (labeled) precision, and (labeled) recall. However, it is well known that these measures do not give an accurate picture of the quality of the parsers output. Furthermore, we will show that they are especially unsuited for partial parsers. In recent years, research has concentrated on dependencybased evaluation measures. We will show in this paper that such a dependency-based evaluation scheme is particularly suitable for partial parsers. TüBa-D, the treebank used here for evaluation, contains all the necessary dependency information so that the conversion of trees into a dependency structure does not have to rely on heuristics. Therefore, the dependency representations are not only reliable, they are also linguistically motivated and can be used for linguistic purposes.
The purpose of this paper is to describe the TüBa-D/Z treebank of written German and to compare it to the independently developed TIGER treebank (Brants et al., 2002). Both treebanks, TIGER and TüBa-D/Z, use an annotation framework that is based on phrase structure grammar and that is enhanced by a level of predicate-argument structure. The comparison between the annotation schemes of the two treebanks focuses on the different treatments of free word order and discontinuous constituents in German as well as on differences in phrase-internal annotation.
The ACL 2008 Workshop on Parsing German features a shared task on parsing German. The goal of the shared task was to find reasons for the radically different behavior of parsers on the different treebanks and between constituent and dependency representations. In this paper, we describe the task and the data sets. In addition, we provide an overview of the test results and a first analysis.
The Conference on Computational Natural Language Learning features a shared task, in which participants train and test their learning systems on the same data sets. In 2007, as in 2006, the shared task has been devoted to dependency parsing, this year with both a multilingual track and a domain adaptation track. In this paper, we define the tasks of the different tracks and describe how the data sets were created from existing treebanks for ten languages. In addition, we characterize the different approaches of the participating systems, report the test results, and provide a first analysis of these results.
Recent approaches to Word Sense Disambiguation (WSD) generally fall into two classes: (1) information-intensive approaches and (2) information-poor approaches. Our hypothesis is that for memory-based learning (MBL), a reduced amount of data is more beneficial than the full range of features used in the past. Our experiments show that MBL combined with a restricted set of features and a feature selection method that minimizes the feature set leads to competitive results, outperforming all systems that participated in the SENSEVAL-3 competition on the Romanian data. Thus, with this specific method, a tightly controlled feature set improves the accuracy of the classifier, reaching 74.0% in the fine-grained and 78.7% in the coarse-grained evaluation.
The purpose of this paper is to describe recent developments in the morphological, syntactic, and semantic annotation of the TüBa-D/Z treebank of German. The TüBa-D/Z annotation scheme is derived from the Verbmobil treebank of spoken German [4, 10], but has been extended along various dimensions to accommodate the characteristics of written texts. TüBa-D/Z uses as its data source the "die tageszeitung" (taz) newspaper corpus. The Verbmobil treebank annotation scheme distinguishes four levels of syntactic constituency: the lexical level, the phrasal level, the level of topological fields, and the clausal level. The primary ordering principle of a clause is the inventory of topological fields, which characterize the word order regularities among different clause types of German, and which are widely accepted among descriptive linguists of German [3, 6]. The TüBa-D/Z annotation relies on a context-free backbone (i.e. proper trees without crossing branches) of phrase structure combined with edge labels that specify the grammatical function of the phrase in question. The syntactic annotation scheme of the TüBa-D/Z is described in more detail in [12, 11]. TüBa-D/Z currently comprises approximately 15 000 sentences, with approximately 7 000 sentences being in the correction phase. The latter will be released along with an updated version of the existing treebank before the end of this year. The treebank is available in an XML format, in the NEGRA export format [1] and in the Penn treebank bracketing format. The XML format contains all types of information as described above, the NEGRA export format contains all sentenceinternal information while the Penn treebank format includes only those layers of information that can be expressed as pure tree structures. Over the course of the last year, more fine grained linguistic annotations have been added along the following dimensions: 1. the basic Stuttgart-Tübingen tagset, STTS, [9] labels have been enriched by relevant features of inflectional morphology, 2. named entity information has been encoded as part of the syntactic annotation, and 3. a set of anaphoric and coreference relations has been added to link referentially dependent noun phrases. In the following sections, we will describe each of these innovations in turn and will demonstrate how the additional annotations can be incorporated into one comprehensive annotation scheme.
The definition of similarity between sentences is formulated on the levels of words, POS tags, and chunks (Abney 91; Abney 96). The evaluation of this approach shows that while precision and recall based on the PARSEVAL measures (Black et al. 91) do not reach state of the art Parsers yet (F1=87.19 on syntactic constituents, F1=77.78 including functionargument structure), the parser shows a very reliable performance where function-argument structure is concerned (F1=96.52). The lower F-scores are very often due to unattached constituents.
The problem of vocalization, or diacritization, is essential to many tasks in Arabic NLP. Arabic is generally written without the short vowels, which leads to one written form having several pronunciations with each pronunciation carrying its own meaning(s). In the experiments reported here, we define vocalization as a classification problem in which we decide for each character in the unvocalized word whether it is followed by a short vowel. We investigate the importance of different types of context. Our results show that the combination of using memory-based learning with only a word internal context leads to a word error rate of 6.64%. If a lexical context is added, the results deteriorate slowly.
In syntax, the trend nowadays is towards lexicalized grammar formalisms. It is now widely accepted that dividing words into wordclasses may serve as a laborsaving mechanism - but at the same time, it discards all detailed information on the idiosyncratic behavior of words. And that is exactly the type of information that may be necessary in order to parse a sentence. For learning approaches, however, lexicalized grammars represent a challenge for the very reason that they include so much detailed and specific information, which is difficult to learn. This paper will present an algorithm for learning a link grammar of German. The problem of data sparseness is tackled by using all the available information from partial parses as well as from an existing grammar fragment and a tagger. This is a report about work in progress so there are no representative results available yet.
This paper presents a comparative study of probabilistic treebank parsing of German, using the Negra and TüBa-D/Z treebanks. Experiments with the Stanford parser, which uses a factored PCFG and dependency model, show that, contrary to previous claims for other parsers, lexicalization of PCFG models boosts parsing performance for both treebanks. The experiments also show that there is a big difference in parsing performance, when trained on the Negra and on the TüBa-D/Z treebanks. Parser performance for the models trained on TüBa-D/Z are comparable to parsing results for English with the Stanford parser, when trained on the Penn treebank. This comparison at least suggests that German is not harder to parse than its West-Germanic neighbor language English.
How to compare treebanks
(2008)
Recent years have seen an increasing interest in developing standards for linguistic annotation, with a focus on the interoperability of the resources. This effort, however, requires a profound knowledge of the advantages and disadvantages of linguistic annotation schemes in order to avoid importing the flaws and weaknesses of existing encoding schemes into the new standards. This paper addresses the question how to compare syntactically annotated corpora and gain insights into the usefulness of specific design decisions. We present an exhaustive evaluation of two German treebanks with crucially different encoding schemes. We evaluate three different parsers trained on the two treebanks and compare results using EVALB, the Leaf-Ancestor metric, and a dependency-based evaluation. Furthermore, we present TePaCoC, a new testsuite for the evaluation of parsers on complex German grammatical constructions. The testsuite provides a well thought-out error classification, which enables us to compare parser output for parsers trained on treebanks with different encoding schemes and provides interesting insights into the impact of treebank annotation schemes on specific constructions like PP attachment or non-constituent coordination.
In the last decade, the Penn treebank has become the standard data set for evaluating parsers. The fact that most parsers are solely evaluated on this specific data set leaves the question unanswered how much these results depend on the annotation scheme of the treebank. In this paper, we will investigate the influence which different decisions in the annotation schemes of treebanks have on parsing. The investigation uses the comparison of similar treebanks of German, NEGRA and TüBa-D/Z, which are subsequently modified to allow a comparison of the differences. The results show that deleted unary nodes and a flat phrase structure have a negative influence on parsing quality while a flat clause structure has a positive influence.
Transforming constituent-based annotation into dependency-based annotation has been shown to work for different treebanks and annotation schemes (e.g. Lin (1995) has transformed the Penn treebank, and Kübler and Telljohann (2002) the Tübinger Baumbank des Deutschen (TüBa-D/Z)). These ventures are usually triggered by the conflict between theory-neutral annotation, that targets most needs of a wider audience, and theory-specific annotation, that provides more fine-grained information for a smaller audience. As a compromise, it has been pointed out that treebanks can be designed to support more than one theory from the start (Nivre, 2003). We argue that information can also be added to an existing annotation scheme so that it supports additional theory-specific annotations. We also argue that such a transformation is useful for improving and extending the original annotation scheme with respect to both ambiguous annotation and annotation errors. We show this by analysing problems that arise when generating dependency information from the constituent-based TüBa-D/Z.
Chunk parsing has focused on the recognition of partial constituent structures at the level of individual chunks. Little attention has been paid to the question of how such partial analyses can be combined into larger structures for complete utterances. Such larger structures are not only desirable for a deeper syntactic analysis. They also constitute a necessary prerequisite for assigning function-argument structure. The present paper offers a similaritybased algorithm for assigning functional labels such as subject, object, head, complement, etc. to complete syntactic structures on the basis of prechunked input. The evaluation of the algorithm has concentrated on measuring the quality of functional labels. It was performed on a German and an English treebank using two different annotation schemes at the level of function argument structure. The results of 89.73% correct functional labels for German and 90.40%for English validate the general approach.
In this paper, we investigate the role of sub-optimality in training data for part-of-speech tagging. In particular, we examine to what extent the size of the training corpus and certain types of errors in it affect the performance of the tagger. We distinguish four types of errors: If a word is assigned a wrong tag, this tag can belong to the ambiguity class of the word (i.e. to the set of possible tags for that word) or not; furthermore, the major syntactic category (e.g. "N" or "V") can be correctly assigned (e.g. if a finite verb is classified as an infinitive) or not (e.g. if a verb is classified as a noun). We empirically explore the decrease of performance that each of these error types causes for different sizes of the training set. Our results show that those types of errors that are easier to eliminate have a particularly negative effect on the performance. Thus, it is worthwhile concentrating on the elimination of these types of errors, especially if the training corpus is large.
Prepositional phrase (PP) attachment is one of the major sources for errors in traditional statistical parsers. The reason for that lies in the type of information necessary for resolving structural ambiguities. For parsing, it is assumed that distributional information of parts-of-speech and phrases is sufficient for disambiguation. For PP attachment, in contrast, lexical information is needed. The problem of PP attachment has sparked much interest ever since Hindle and Rooth (1993) formulated the problem in a way that can be easily handled by machine learning approaches: In their approach, PP attachment is reduced to the decision between noun and verb attachment; and the relevant information is reduced to the two possible attachment sites (the noun and the verb) and the preposition of the PP. Brill and Resnik (1994) extended the feature set to the now standard 4-tupel also containing the noun inside the PP. Among many publications on the problem of PP attachment, Volk (2001; 2002) describes the only system for German. He uses a combination of supervised and unsupervised methods. The supervised method is based on the back-off model by Collins and Brooks (1995), the unsupervised part consists of heuristics such as ”If there is a support verb construction present, choose verb attachment”. Volk trains his back-off model on the Negra treebank (Skut et al., 1998) and extracts frequencies for the heuristics from the ”Computerzeitung”. The latter also serves as test data set. Consequently, it is difficult to compare Volk’s results to other results for German, including the results presented here, since not only he uses a combination of supervised and unsupervised learning, but he also performs domain adaptation. Most of the researchers working on PP attachment seem to be satisfied with a PP attachment system; we have found hardly any work on integrating the results of such approaches into actual parsers. The only exceptions are Mehl et al. (1998) and Foth and Menzel (2006), both working with German data. Mehl et al. report a slight improvement of PP attachment from 475 correct PPs out of 681 PPs for the original parser to 481 PPs. Foth and Menzel report an improvement of overall accuracy from 90.7% to 92.2%. Both integrate statistical attachment preferences into a parser. First, we will investigate whether dependency parsing, which generally uses lexical information, shows the same performance on PP attachment as an independent PP attachment classifier does. Then we will investigate an approach that allows the integration of PP attachment information into the output of a parser without having to modify the parser: The results of an independent PP attachment classifier are integrated into the parse of a dependency parser for German in a postprocessing step.
This report explores the question of compatibility between annotation projects including translating annotation formalisms to each other or to common forms. Compatibility issues are crucial for systems that use the results of multiple annotation projects. We hope that this report will begin a concerted effort in the field to track the compatibility of annotation schemes for part of speech tagging, time annotation, treebanking, role labeling and other phenomena.
This paper reports on the SYN-RA (SYNtax-based Reference Annotation) project, an on-going project of annotating German newspaper texts with referential relations. The project has developed an inventory of anaphoric and coreference relations for German in the context of a unified, XML-based annotation scheme for combining morphological, syntactic, semantic, and anaphoric information. The paper discusses how this unified annotation scheme relates to other formats currently discussed in the literature, in particular the annotation graph model of Bird and Liberman (2001) and the pie-in-thesky scheme for semantic annotation.
Chunk parsing has focused on the recognition of partial constituent structures at the level of individual chunks. Little attention has been paid to the question of how such partial analyses can be combined into larger structures for complete utterances. The TüSBL parser extends current chunk parsing techniques by a tree-construction component that extends partial chunk parses to complete tree structures including recursive phrase structure as well as function-argument structure. TüSBLs tree construction algorithm relies on techniques from memory-based learning that allow similarity-based classification of a given input structure relative to a pre-stored set of tree instances from a fully annotated treebank. A quantitative evaluation of TüSBL has been conducted using a semi-automatically constructed treebank of German that consists of appr. 67,000 fully annotated sentences. The basic PARSEVAL measures were used although they were developed for parsers that have as their main goal a complete analysis that spans the entire input.This runs counter to the basic philosophy underlying TüSBL, which has as its main goal robustness of partially analyzed structures.
This paper provides an overview of current research on a hybrid and robust parsing architecture for the morphological, syntactic and semantic annotation of German text corpora. The novel contribution of this research lies not in the individual parsing modules, each of which relies on state-of-the-art algorithms and techniques. Rather what is new about the present approach is the combination of these modules into a single architecture. This combination provides a means to significantly optimize the performance of each component, resulting in an increased accuracy of annotation.
A lot of interest has recently been paid to constraint-based definitions and extensions of Tree Adjoining Grammars (TAG). Examples are the so-called quasi-trees, D-Tree Grammars and Tree Description Grammars. The latter are grammars consisting of a set of formulars denoting trees. TDGs are derivation based where in each derivation step a conjunction is built of the old formular, a formular of the grammar and additional equivalences between node names of the two formulars. This formalism is more powerfull than TAGs. TDGs offer the advantages of MC-TAG and D-Tree Grammars for natural languages and they allow underspecification. However the problem is that TDGs might be unnecessarily powerfull for natural languages. To solve this problem, in this paper, I will propose a local TDGs, a restricted version of TDGs. Local TDGs still have the advantages of TDGs but they are semilinear and therefore more appropriate for natural languages. First, the notion of the semilinearity is defined. Then local TDGs are introduced, and, finally, semilinearity of local Tree Description Languages is proven.
This paper proposes a compositional semantics for lexicalized tree adjoining grammars (LTAG). Tree-local multicompnent derivations allow seperation of semantiv contribution of a lexical item into one component contributing to the predicate argument structure and second a component contributing to scope semantics. Based on this idea a syntx-semantics interface is presented where the compositional semantics depends only on the derivation structure. It is shown that the derivation structure allows an appropriate amount of underspecification. This is illustrated by investigating underspecified representations for quantifier scpoe ambiguities and related phenomena such as adjunct scope and island constraints.