Cassia Trojahn

Also published as: Cássia Trojahn


Extracting hypernym relations from Wikipedia disambiguation pages : comparing symbolic and machine learning approaches
Mouna Kamel | Cassia Trojahn | Adel Ghamnia | Nathalie Aussenac-Gilles | Cécile Fabre
IWCS 2017 - 12th International Conference on Computational Semantics - Long papers


VOAR: A Visual and Integrated Ontology Alignment Environment
Bernardo Severo | Cassia Trojahn | Renata Vieira
Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)

Ontology alignment is a key process for enabling interoperability between ontology-based systems in the Linked Open Data age. From two input ontologies, this process generates an alignment (set of correspondences) between them. In this paper we present VOAR, a new web-based environment for ontology alignment visualization and manipulation. Within this graphical environment, users can manually create/edit correspondences and apply a set of operations on alignments (filtering, merge, difference, etc.). VOAR allows invoking external ontology matching systems that implement a specific alignment interface, so that the generated alignments can be manipulated within the environment. Evaluating multiple alignments together against a reference one can also be carried out, using classical evaluation metrics (precision, recall and f-measure). The status of each correspondence with respect to its presence or absence in reference alignment is visually represented. Overall, the main new aspect of VOAR is the visualization and manipulation of alignments at schema level, in an integrated, visual and web-based environment.


A Comparable Corpus Based on Aligned Multilingual Ontologies
Roger Granada | Lucelene Lopes | Carlos Ramisch | Cassia Trojahn | Renata Vieira | Aline Villavicencio
Proceedings of the First Workshop on Multilingual Modeling


An API for Multi-lingual Ontology Matching
Cássia Trojahn | Paulo Quaresma | Renata Vieira
Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)

Ontology matching consists of generating a set of correspondences between the entities of two ontologies. This process is seen as a solution to data heterogeneity in ontology-based applications, enabling the interoperability between them. However, existing matching systems are designed by assuming that the entities of both source and target ontologies are written in the same languages ( English, for instance). Multi-lingual ontology matching is an open research issue. This paper describes an API for multi-lingual matching that implements two strategies, direct translation-based and indirect. The first strategy considers direct matching between two ontologies (i.e., without intermediary ontologies), with the help of external resources, i.e., translations. The indirect alignment strategy, proposed by (Jung et al., 2009), is based on composition of alignments. We evaluate these strategies using simple string similarity based matchers and three ontologies written in English, French, and Portuguese, an extension of the OAEI benchmark test 206.


A Framework for Multilingual Ontology Mapping
Cássia Trojahn | Paulo Quaresma | Renata Vieira
Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)

In the field of ontology mapping, multilingual ontology mapping is an issue that is not well explored. This paper proposes a framework for mapping of multilingual Description Logics (DL) ontologies. First, the DL source ontology is translated to the target ontology language, using a lexical database or a dictionary, generating a DL translated ontology. The target and the translated ontologies are then used as input for the mapping process. The mappings are computed by specialized agents using different mapping approaches. Next, these agents use argumentation to exchange their local results, in order to agree on the obtained mappings. Based on their preferences and confidence of the arguments, the agents compute their preferred mapping sets. The arguments in such preferred sets are viewed as the set of globally acceptable arguments. A DL mapping ontology is generated as result of the mapping process. In this paper we focus on the process of generating the DL translated ontology.