Refine
Document Type
- Conference Proceeding (18)
- Article (3)
- Research Data (1)
- Working Paper (1)
Keywords
The OntoLex vocabulary has become a widely used community standard for machine-readable lexical resources on the web. The primary motivation to use OntoLex in favor of tool- or application-specific formalisms is to facilitate interoperability and information integration across different resources. One of its extension that is currently being developed is a module for representing morphology, OntoLex-Morph. In this paper, we show how OntoLex-Morph can be used for the encoding and integration of different types of morphological resources on a unified basis. With German as the example, we demonstrate it for (a) a full-form dictionary with inflection information (Unimorph), (b) a dictionary of base forms and their derivations (UDer), (c) a dictionary of compounds (from GermaNet), and (d) lexicon and inflection rules of a finite-state parser/generator (SMOR/Morphisto). These data are converted to OntoLex-Morph, their linguistic information is consolidated and corresponding lexical entries are linked with each other. The main contribution of this paper is the discussion of the current state of OntoLex-Morph and its validation on different types of real-world resources for a single language. In the longer term, the successful application of OntoLex-Morph to such diverse data, along with the adjustments to the vocabulary observed in the process, will be a means to establish interoperability among morphological resources as well as between them and classical lexical data such as dictionaries, WordNets, or thesauri.
Following presentations of frequency and attestations, and embeddings and distributional similarity, this paper introduces the third cornerstone of the emerging OntoLex module for Frequency, Attestation and Corpus-based Information, OntoLex-FrAC. We provide an RDF vocabulary for collocations, established as a consensus over contributions from five different institutions and numerous data sets, with the goal of eliciting feedback from reviewers, workshop audience and the scientific community in preparation of the final consolidation of the OntoLex-FrAC module, whose publication as a W3C community report is foreseen for the end of this year. The novel collocation component of OntoLex-FrAC is described in application to a lexicographic resource and corpus-based collocation scores available from the web, and finally, we demonstrate the capability and genericity of the model by showing how to retrieve and aggregate collocation information by means of SPARQL, and its export to a tabular format, so that it can be easily processed in downstream applications.
The OntoLex vocabulary enjoys increasing popularity as a means of publishing lexical resources with RDF and as Linked Data. The
recent publication of a new OntoLex module for lexicography, lexicog, reflects its increasing importance for digital lexicography.
However, not all aspects of digital lexicography have been covered to the same extent. In particular, supplementary information drawn
from corpora such as frequency information, links to attestations, and collocation data were considered to be beyond the scope of
lexicog. Therefore, the OntoLex community has put forward the proposal for a novel module for frequency, attestation and corpus
information (FrAC), that not only covers the requirements of digital lexicography, but also accommodates essential data structures for
lexical information in natural language processing. This paper introduces the current state of the OntoLex-FrAC vocabulary, describes
its structure, some selected use cases, elementary concepts and fundamental definitions, with a focus on frequency and attestations.
The paper describes the first comprehensive edition of machine-readable discourse marker lexicons. Discourse markers such as and, because, but, though or thereafter are essential communicative signals in human conversation, as they indicate how an utterance relates to its communicative context. As much of this information is implicit or expressed differently in different languages, discourse parsing, context-adequate natural language generation and machine translation are considered particularly challenging aspects of Natural Language Processing. Providing this data in machine-readable, standard-compliant form will thus facilitate such technical tasks, and moreover, allow to explore techniques for translation inference to be applied to this particular group of lexical resources that was previously largely neglected in the context of Linguistic Linked (Open) Data.
This article provides a comprehensive and up-to-date survey of models and vocabularies for creating linguistic linked data (LLD) focusing on the latest developments in the area and both building upon and complementing previous works covering similar territory. The article begins with an overview of some recent trends which have had a significant impact on linked data models and vocabularies. Next, we give a general overview of existing vocabularies and models for different categories of LLD resource. After which we look at some of the latest developments in community standards and initiatives including descriptions of recent work on the OntoLex-Lemon model, a survey of recent initiatives in linguistic annotation and LLD, and a discussion of the LLD metadata vocabularies META-SHARE and lime. In the next part of the paper, we focus on the influence of projects on LLD models and vocabularies, starting with a general survey of relevant projects, before dedicating individual sections to a number of recent projects and their impact on LLD vocabularies and models. Finally, in the conclusion, we look ahead at some future challenges for LLD models and vocabularies. The appendix to the paper consists of a brief introduction to the OntoLex-Lemon model.
In language technology and language sciences, tab-separated values (TSV) represent a frequently used formalism to represent linguistically annotated natural language, often addressed as "CoNLL formats". A large number of such formats do exist, but although they share a number of common features, they are not interoperable, as different pieces of information are encoded differently in these dialects.
CoNLL-RDF refers to a programming library and the associated data model that has been introduced to facilitate processing and transforming such TSV formats in a serialization-independent way. CoNLL-RDF represents CoNLL data, by means of RDF graphs and SPARQL update operations, but so far, without machine-readable semantics, with annotation properties created dynamically on the basis of a user-defined mapping from columns to labels. Current applications of CoNLL-RDF include linking between corpora and dictionaries [Mambrini and Passarotti, 2019] and knowledge graphs [Tamper et al., 2018], syntactic parsing of historical languages [Chiarcos et al., 2018; Chiarcos et al., 2018], the consolidation of syntactic and semantic annotations [Chiarcos and Fäth, 2019], a bridge between RDF corpora and a traditional corpus query language [Ionov et al., 2020], and language contact studies [Chiarcos et al., 2018].
We describe a novel extension of CoNLL-RDF, introducing a formal data model, formalized as an ontology. The ontology is a basis for linking RDF corpora with other Semantic Web resources, but more importantly, its application for transformation between different TSV formats is a major step for providing interoperability between CoNLL formats.
Large-scale diachronic corpus studies covering longer time periods are difficult if more than one corpus are to be consulted and, as a result, different formats and annotation schemas need to be processed and queried in a uniform, comparable and replicable manner. We describes the application of the Flexible Integrated Transformation and Annotation eNgineering (Fintan) platform for studying word order in German using syntactically annotated corpora that represent its entire written history. Focusing on nominal dative and accusative arguments, this study hints at two major phases in the development of scrambling in modern German. Against more recent assumptions, it supports the traditional view that word order flexibility decreased over time, but it also indicates that this was a relatively sharp transition in Early New High German. The successful case study demonstrates the potential of Fintan and the underlying LLOD technology for historical linguistics, linguistic typology and corpus linguistics. The technological contribution of this paper is to demonstrate the applicability of Fintan for querying across heterogeneously annotated corpora, as previously, it had only been applied for transformation tasks. With its focus on quantitative analysis, Fintan is a natural complement for existing multi-layer technologies that focus on query and exploration.
The Caucasus region is famed for its rich and diverse arrays of languages and language families, often challenging European-centered
views established in traditional linguistics. In this paper, we describe ongoing efforts to improve the coverage of Universal Morphologies for languages of the Caucasus region. The Universal Morphologies (UniMorph) are a recent community project aiming to complement the Universal Dependencies which focus on morphosyntax and syntax. We describe the development of UniMorph resources for
Nakh-Daghestanian and Kartvelian languages as a well as for Classical Armenian, we discuss challenges that the complex morphology of these and related languages poses to the current design of UniMorph, and suggest possibilities to improve the applicability of UniMorph for languages of the Caucasus region in particular and for low resource languages in general. We also criticize the UniMorph TSV format for its limited expressiveness, and suggest to complement the existing UniMorph workflow with support for additional source formats on grounds of Linked Open Data technology.
We introduce the Flexible and Integrated Transformation and Annotation eNgeneering (Fintan) platform for converting heterogeneous linguistic resources to RDF. With its modular architecture, workflow management and visualization features, Fintan facilitates the development of complex transformation pipelines by integrating generic RDF converters and augmenting them with extended graph processing capabilities: Existing converters can be easily deployed to the system by means of an ontological data structure which renders their properties and the dependencies between transformation steps. Development of subsequent graph transformation steps for resource transformation, annotation engineering or entity linking is further facilitated by a novel visual rendering of SPARQL queries. A graphical workflow manager allows to easily manage the converter modules and combine them to new transformation pipelines. Employing the stream-based graph processing approach first implemented with CoNLL-RDF, we address common challenges and scalability issues when transforming resources and showcase the performance of Fintan by means of a purely graph-based transformation of the Universal Morphology data to RDF.
The ACoLi dictionary graph
(2020)
In this paper, we report the release of the ACoLi Dictionary Graph, a large-scale collection of multilingual open source dictionaries available in two machine-readable formats, a graph representation in RDF, using the OntoLex-Lemon vocabulary, and a simple tabular data format to facilitate their use in NLP tasks, such as translation inference across dictionaries. We describe the mapping and harmonization of the underlying data structures into a unified representation, its serialization in RDF and TSV, and the release of a massive and coherent amount of lexical data under open licenses.
Recent years have experienced a growing trend in the publication of language resources as Linguistic Linked Data (LLD) to enhance their discovery, reuse and the interoperability of tools that consume language data. To this aim, the OntoLex-lemon model has emerged as a de facto standard to represent lexical data on the Web. However, traditional dictionaries contain a considerable amount of morphological information which is not straightforwardly representable
as LLD within the current model. In order to fill this gap a new Morphology Module of OntoLex-lemon is currently being developed. This paper presents the results of this model as on-going work as well as the underlying challenges that emerged during the module
development. Based on the MMoOn Core ontology, it aims to account for a wide range of morphological information, ranging from endings to derive whole paradigms to the decomposition and generation of lexical entries which is in compliance to other OntoLex-lemon modules and facilitates the encoding of complex morphological data in ontology lexicons.
The paper introduces Ligt, a native RDF vocabulary for representing linguistic examples as text with interlinear glosses (IGT) in a linked data formalism. Interlinear glossing is a notation used in various fields of linguistics to provide readers with a way to understand linguistic phenomena and to provide corpus data when documenting endangered languages. This data is usually provided with morpheme-by-morpheme correspondence which is not supported by any established vocabularies for representing linguistic corpora or automated annotations. Interlinear Glossed Text can be stored and exchanged in several formats specifically designed for the purpose, but these differ in their designs and concepts, and they are tied to particular tools, so the reusability of the annotated data is limited. To improve interoperability and reusability, we propose to convert such glosses to a tool-independent representation well-suited for the Web of Data, i.e., a representation in RDF. Beyond establishing structural (format) interoperability by means of a common data representation, our approach also allows using shared vocabularies and terminology repositories available from the (Linguistic) Linked Open Data cloud.
We describe the core vocabulary and the converters that use this vocabulary to convert IGT in a format of various widely-used tools into RDF. Ultimately, a Linked Data representation will facilitate the accessibility of language data from less-resourced language varieties within the (Linguistic) Linked Open Data cloud, as well as enable novel ways to access and integrate this information with (L)LOD dictionary data and other types of lexical-semantic resources. In a longer perspective, data currently only available through these formats will become more visible and reusable and contribute to the development of a truly multilingual (semantic) web.
This paper describes our contribution to the closed track of the Shared Task Translation Inference across Dictionaries (TIAD2017), 1 held in conjunction with the first Conference on Language Data and Knowledge (LDK-2017). In our approach, we use supervised machine learning to predict high-quality candidate translation pairs. We train a Support Vector Machine using several features, mostly of the translation graph, but also taking into consideration string similarity (Levenshtein
distance). As the closed track does not provide manual training data, we
define positive training examples as translation candidate pairs which
occur in a cycle in which there is a direct connection.
