A Data-Driven Dependency Parser for Bulgarian

2014

In this paper we present a system for experimenting with combinations of dependency parsers. The system supports initial training of different parsing models, creation of parsebank(s) with these models, and different strategies for the construction of ensemble models aimed at improving the output of the individual models by voting. The system employs two algorithms for construction of dependency trees from several parses of the same sentence and several ways for ranking of the arcs in the resulting trees. We have performed experiments with state-of-the-art dependency parsers including MaltParser, MSTParser, TurboParser, and MATEParser, on the data from the Bulgarian treebank -- BulTreeBank. Our best result from these experiments is slightly better then the best result reported in the literature for this language.

In this paper, we analyze the impact of various dependency representations for various constructions on the general parsing accuracy and on the parsing accuracy of these constructions. We focus on the analysis of coordination constructions, complex predicates, and punctuation mark attachment. We use Latvian Treebank as a dataset, thus, providing insight for an inflective language with a rather free word order. Experiments with MaltParser, a transition-based parser, show clear difference in learnability of various representations for the considered constructions. Future work would include carrying out comparable experiments with a graph-based dependency parser like MSTParser.

We present the results of the experiment of bootstrapping a Treebank for Catalan by using a Dependency Parser trained with Spanish sentences. In order to save time and cost, our approach was to profit from the typological similarities between Catalan and Spanish to create a first Catalan data set quickly by (i) automatically annotating with a delexicalized Spanish parser, (ii) manually correcting the parses, and (iii) using the Catalan corrected sentences to train a Catalan parser. The results showed that the number of parsed sentences required to train a Catalan parser is about 1000, which were achieved in 4 months with 2 annotators.

Proceedings of the, 2018

We present the Uppsala system for the CoNLL 2018 Shared Task on universal dependency parsing. Our system is a pipeline consisting of three components: the first performs joint word and sentence segmentation; the second predicts part-ofspeech tags and morphological features; the third predicts dependency trees from words and tags. Instead of training a single parsing model for each treebank, we trained models with multiple treebanks for one language or closely related languages, greatly reducing the number of models. On the official test run, we ranked 7th of 27 teams for the LAS and MLAS metrics. Our system obtained the best scores overall for word segmentation, universal POS tagging, and morphological features. 2 Resources All three components of our system were trained principally on the training sets of Universal Dependencies v2.2 released to coincide with the shared task (Nivre et al., 2018). The tagger and parser also make use of the pre-trained word

2003

Many extensions to text-based, data-intensive knowledge management approaches, such as Information Retrieval or Data Mining, focus on integrating the impressive recent advances in language technology. For this, they need fast, robust parsers that deliver linguistic data which is meaningful for the subsequent processing stages. This paper introduces such a parsing system. Its output is a hierarchical structure of syntactic relations, functional dependency structures.

We present the current state of development of the Croatian Dependency Treebank -with special empahsis on adapting the Prague Dependency Treebank formalism to Croatian language specifics -and illustrate its possible applications in an experiment with dependency parsing using MaltParser. The treebank currently contains approximately 2870 sentences, out of which the 2699 sentences and 66930 tokens were used in this experiment. Three linear-time projective algorithms implemented by the MaltParser system -Nivre eager, Nivre standard and stack projective -running on default settings were used in the experiment. The highest performing system, implementing the Nivre eager algorithm, scored (LAS 71.31 UAS 80.93 LA 83.87) within our experiment setup. The results obtained serve as an illustration of treebank's usefulness in natural language processing research and as a baseline for further research in dependency parsing of Croatian.

proceedings of the …, 2006

Recently dependency parsing is gaining popularity. It is broadly accepted that dependency representations are more suitable for free word order languages. Statistical dependency parsers are easy to port from one language to another, if there are dependency treebanks for learning a grammar for the particular language. However, many treebanks are based on constituency and have to be converted to dependency representations prior to learning statistical dependency parsers. In this paper we investigate the issues of the conversion of the BulTreeBank (Simov et al., 2002) from Head-driven Phrase Structure Grammar (HPSG) format to dependency-based format and its parsing. We have performed three different conversions to three different dependency formats. For two of the conversions we used head tables and dependency tables which were stated explicitly, as in (Xia, 2001). For the other conversion the tables were implicitly implemented by rules. Our choice of rules for the tables was guided by decisions rooted in different linguistic theories. We have parsed the converted treebank with the Malt parser (Nivre et al., 2004) for 'evaluating' our conversions. Then we made error analysis to find advantages and pitfalls of each conversion strategy.

Ninth International Workshop on Treebanks …, 2010

The Russian syntactic treebank SynTagRus is annotated with dependency structures in line with the Meaning-Text Theory (MTT). In order to benefit from the detailed syntactic annotation in SynTagRus and facilitate the development of a Russian Resource Grammar (RRG) in the framework of Head-driven Phrase Structure Grammar (HPSG), we need to convert the dependency structures into HPSG derivation trees. Our pilot study has shown that many of the constructions can be converted systematically with simple rules. In order to extend the depth and coverage of this conversion, we need to implement conversion heuristics that produce linguistically sound HPSG derivations. As a result we obtain a structured set of correspondences between MTT surface syntactic relations and HPSG phrasal types, which enable the cross-theoretical transfer of insightful syntactic analyses and formalized deep linguistic knowledge. The converted treebank SynTagRus++ is annotated with HPSG structures and of crucial importance to the RRG under development, as our goal is to ensure an optimal and efficient grammar engineering cycle through dynamic coupling of the treebank and the grammar. * We are grateful to Leonid L. Iomdin for providing us with access to the SynTagRus dependency treebank and for helpful answers to annotation-related questions.

2004

Abstract This paper describes a method for conducting evaluations of Treebank and non-Treebank parsers alike against the English language U. Penn Treebank (Marcus et al., 1993) using a metric that focuses on the accuracy of relatively non-controversial aspects of parse structure. Our conjecture is that if we focus on maximal projections of heads (MPH), we are likely to find much broader agreement than if we try to evaluate based on order of attachment.

Proceedings of the Workshop on Computational …, 1997

I describe the TreeBanker, a graphical tool for the supervised training involved in domain customization of the disambiguation component of a speech-or languageunderstanding system. The TreeBanker presents a user, who need not be a ...

2005

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, NE-GRA 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.

Software Engineering, Testing, and Quality Assurance for Natural Language Processing on - SETQA-NLP '08, 2008

Natural language processing modules such as part-of-speech taggers, named-entity recognizers and syntactic parsers are commonly evaluated in isolation, under the assumption that artificial evaluation metrics for individual parts are predictive of practical performance of more complex language technology systems that perform practical tasks. Although this is an important issue in the design and engineering of systems that use natural language input, it is often unclear how the accuracy of an end-user application is affected by parameters that affect individual NLP modules. We explore this issue in the context of a specific task by examining the relationship between the accuracy of a syntactic parser and the overall performance of an information extraction system for biomedical text that includes the parser as one of its components. We present an empirical investigation of the relationship between factors that affect the accuracy of syntactic analysis, and how the difference in parse accuracy affects the overall system.

2010

We first describe the automatic conversion of the French Treebank (Abeillé and , a constituency treebank, into typed projective dependency trees. In order to evaluate the overall quality of the resulting dependency treebank, and to quantify the cases where the projectivity constraint leads to wrong dependencies, we compare a subset of the converted treebank to manually validated dependency trees. We then compare the performance of two treebank-trained parsers that output typed dependency parses. The first parser is the MST parser , which we directly train on dependency trees. The second parser is a combination of the Berkeley parser (Petrov et al., 2006) and a functional role labeler: trained on the original constituency treebank, the Berkeley parser first outputs constituency trees, which are then labeled with functional roles, and then converted into dependency trees. We found that used in combination with a high-accuracy French POS tagger, the MST parser performs a little better for unlabeled dependencies (UAS=90.3% versus 89.6%), and better for labeled dependencies (LAS=87.6% versus 85.6%).

2007

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. 2 Task Definition In this section, we provide the task definitions that were used in the two tracks of the CoNLL 2007 Shard Task, the multilingual track and the domain adaptation track, together with some background and motivation for the design choices made. First of all, we give a brief description of the data format and evaluation metrics, which were common to the two tracks. 2.1 Data Format and Evaluation Metrics

We present a new version of the Croatian Dependency Treebank. It constitutes a slight departure from the previously closely observed Prague Dependency Treebank syntactic layer annotation guidelines as we introduce a new subset of syntactic tags on top of the existing tagset. These new tags are used in explicit annotation of subordinate clauses via subordinate conjunctions. Introducing the new annotation to Croatian Dependency Treebank, we also modify head attachment rules addressing subordinate conjunctions and subordinate clause predicates. In an experiment with data-driven dependency parsing, we show that implementing these new annotation guidelines leeds to a statistically significant improvement in parsing accuracy. We also observe a substantial improvement in inter-annotator agreement, facilitating more consistent annotation in further treebank development.