Thierry Declerck

The automatic generation of formal annotations in a multimedia indexing and searching environment

We describe in this paper the MU-MIS Project (Multimedia Indexing and Searching Environment), which is concerned with the development and integration of base technologies, demonstrated within a laboratory prototype, to support automated multimedia indexing and to facilitate search and retrieval from multimedia databases. We stress the role linguistically motivated annotations, coupled with domain-specific information, can play within this environment. The project will demonstrate that innovative technology components can operate on multilingual, multisource, and multimedia information and create a meaningful and queryable database.

Language Service Ontology

The Language Grid is a distinctive language service infrastructure in the sense that it accommodates a wide variety of user needs, ranging from technical novices to experts; language resource consumers to language resource providers. As these language services are various in type and each of them can be idiosyncratic in many aspects, the service infrastructure has to address the issue of interoperability. A key to solve this issue is not only to build the services around standardized resources and interfaces, but also to establish a knowledge structure that copes effectively with a range of language services. Given this knowledge structure, referred to as a service ontology, each language service can be systematically classified and its usage specified by a corresponding API. This not only enables the utilization of existing language resources but facilitates the dissemination of newly created language resources as services.

How to semantically relate dialectal Dictionaries in the Linked Data Framework

We describe on-going work towards publishing language resources included in dialectal dictionaries in the Linked Open Data (LOD) cloud, and so to support wider access to the diverse cultural data associated with such dictionary entries, like the various historical and geographical variations of the use of such words. Beyond this, our approach allows the cross-linking of entries of dialectal dictionaries on the basis of the semantic representation of their senses, and also to link the entries of the dialectal dictionaries to lexical senses available in the LOD framework. This paper focuses on the description of the steps leading to a SKOS-XL and lemon encoding of the entries of two Austrian dialectal dictionaries, and how this work supports their cross-linking and linking to other language data in the LOD.

Towards Linked Language Data for Digital Humanities

We investigate the extension of classification schemes in the Humanities into semantic data repositories, the benefits of which could be the automation of so far manually conducted processes, such as detecting motifs in folktale texts. In parallel, we propose linguistic analysis of the textual labels used in these repositories. The resulting resource, which we propose to publish in the Linked Open Data (LOD) framework, will explicitly interlink domain knowledge and linguistically enriched language data, which can be used for knowledge-driven content analysis of literary works.

Proppian Content Descriptors in an Integrated Annotation Schema for Fairy Tales

This chapter describes the actual state of APftML (Augmented Proppian fairy tale Markup Language), which is a schema combining linguistic and domain specific annotation for supporting Cultural Heritage and Digital Humanities research, exemplified in the fairy tale domain. APftML should in particular guide automated text analysis to detect and mark up fairy tale characters and the typicalactions they are involved in, which can be subsequently queried in a corpus by both linguists and specialists in the field. The characters and actions are defined with the help of Propp’s formal analysis of folktales, which we aim to implement in a fully fledged way, contrary to existing computational resources based on his theory. Inorder to respond to current formalisation requirements APfML abstracts away from some aspects of the theory of Propp and we also discuss the integration of Proppian elements within modern semantic annotation approaches. The chapter focus on the resulting revised and extended set of narrative elements APftML is dealing with.

Event-coreference across multiple, multi-lingual sources in the Mumis project

We present our work on information extraction from multiple, multi-lingual sources for the Multimedia Indexing and Searching Environment (MUMIS), a project aiming at developing technology to produce formal annotations about essential events in multimedia programme material. The novelty of our approach consists on the use of a merging or cross-document coreference algorithm that aims at combining the output delivered by the information extraction systems.

SentiMerge: Combining Sentiment Lexicons in a Bayesian Framework

Many approaches to sentiment analysis rely on a lexicon that labels words with a prior polarity. This is particularly true for languages other than English, where labelled training data is not easily available. Existing efforts to produce such lexicons exist, and to avoid duplicated effort, a principled way to combine multiple resources is required. In this paper, we introduce a Bayesian probabilistic model, which can simultaneously combine polarity scores from several data sources and estimate the quality of each source. We apply this algorithm to a set of four German sentiment lexicons, to produce the SentiMerge lexicon, which we make publically available. In a simple classification task, we show that this lexicon outperforms each of the underlying resources, as well as a majority vote model.

Concept and Relation Extraction in the Finance Domain

In this paper, we describe the state of our work on the possible derivation of ontological structures from textual analysis. We propose an approach to semi-automatic generation of domain ontologies from scratch, on the basis of heuristic rules applied to the result of a multi-layered processing of textual documents.

A Hybrid Reasoning Architecture for Business Intelligence Applications

We describe an implemented hybrid reasoning architecture that is used in an EU-funded project called MUSING (www.musing.eu) which is dedicated towards the investigation of semantic-based business intelligence solutions. The reasoning platform builds on publicly available software, such as Pellet, OWLIM, Jena, and Sesame. The project uses and extends existing OWL ontologies (e.g., PROTON) and assumes rule-based reasoning to take place on top of OWL. We describe the pros and cons of each subsystem w.r.t. the needs we have encountered during our investigation. We explain the specific reasoning architecture that is based on a sequence and a fixpoint computation of three reasoners which we might sloppily write as Pellet + (OWLIM + Jena)^*. Pellet is used for checking the initial consistency of the ontology, whereas OWLIM and Jena are employed to execute rules outside the expressiveness of OWL. However, OWLIM is way much faster than Jena, but neither has means to do numerical comparison nor arithmetic. We explain our choice why SWRL is not enough and why we believe that binary OWL properties lead to an unwanted proliferation of objects, making representation and reasoning extremely complex.

Cross-media semantic indexing in the soccer domain

In this paper we describe collaborative and integrative work in the K-Space Network of Excellence. A goal of the work presented consists of combining results of the analysis of soccer videos with the semantic analysis of textual complementary sources, in order to support the semantic annotation and indexing of soccer videos. We present briefly a former approach to text-based semantic annotation and indexing of soccer videos as done in the MUMIS project and show in comparison the advances achieved within the K-Space project. A SMIL-based demonstrator has been implemented, documenting the approach and the resources used in our work. We present then briefly the set up of this demonstrator.