In 2004 we published in this journal an article describing OntoLearn, one of the first systems to automatically induce a taxonomy from documents and Web sites. Since then, OntoLearn has continued be active area research our group become reference work within community. paper describe next-generation learning methodology, which name Reloaded. Unlike many approaches literature, novel algorithm learns both concepts relations entirely scratch via automated extraction terms, definitions,...

In this paper we present a graph-based approach aimed at learning lexical taxonomy automatically starting from domain corpus and the Web. Unlike many approaches in literature, our novel algorithm learns both concepts relations entirely scratch via automated extraction of terms, definitions hypernyms. This results very dense, cyclic possibly disconnected hypernym graph. The then induces Our experiments show that obtain high-quality results, when building brand-new taxonomies reconstructing...

This paper describes the first shared task on Taxonomy Extraction Evaluation organised as part of SemEval-2015.Participants were asked to find hypernym-hyponym relations between given terms.For each four selected target domains participants provided with two lists domainspecific terms: a WordNet collection terms and well-known terminology extracted from an online publicly available taxonomy.A total 45 taxonomies submitted by 6 participating teams evaluated using standard structural measures,...

Alexander Panchenko, Stefano Faralli, Eugen Ruppert, Steffen Remus, Hubert Naets, Cédrick Fairon, Simone Paolo Ponzetto, Chris Biemann. Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016). 2016.

In this paper we present a novel approach to learning semantic models for multiple domains, which use categorize Wikipedia pages and perform domain Word Sense Disambiguation (WSD). order learn model each first extract relevant terms from the texts in then these initialize random walk over WordNet graph. Given an input text, check models, choose appropriate that text best-matching WSD. Our results show considerable improvements on categorization WSD tasks.

Alexander Panchenko, Eugen Ruppert, Stefano Faralli, Simone Paolo Ponzetto, Chris Biemann. Proceedings of the 15th Conference European Chapter Association for Computational Linguistics: Volume 1, Long Papers. 2017.

Individuals, local communities, environmental associations, private organizations, and public representatives bodies may all be aggrieved by problems concerning poor air quality, illegal waste disposal, water contamination, general pollution. Environmental complaints represent the expressions of dissatisfaction with these issues. As time-consuming managing a large number complaints, text mining useful for automatically extracting information on stakeholder priorities concerns. The paper used...

Stefano Faralli, Alexander Panchenko, Chris Biemann, Simone Paolo Ponzetto. Proceedings of the 15th Conference European Chapter Association for Computational Linguistics: Volume 1, Long Papers. 2017.

Taxonomies are an important ingredient of knowledge organization, and serve as a backbone for more sophisticated representations in intelligent systems, such formal ontologies. However, building taxonomies manually is costly endeavor, hence, automatic methods taxonomy induction good alternative to build large-scale taxonomies. In this paper, we propose TIEmb, approach unsupervised class subsumption axiom extraction from bases using entity text embeddings. We apply the on WebIsA database,...

Interpretability of a predictive model is powerful feature that gains the trust users in correctness predictions. In word sense disambiguation (WSD), knowledge-based systems tend to be much more interpretable than knowledge-free counterparts as they rely on wealth manually-encoded elements representing senses, such hypernyms, usage examples, and images. We present WSD system bridges gap between these two so far disconnected groups methods. Namely, our system, providing access several...

Abstract We present an approach to combining distributional semantic representations induced from text corpora with manually constructed lexical networks. While both kinds of resources are available high coverage, our aligned resource combines the domain specificity and availability contextual information models conciseness quality crafted start a representation senses vocabulary terms, which accompanied rich context given by related items. then automatically disambiguate such obtain...

This paper presents MILA, a prototype interactive Learning Analytics tool for the Moodle learning management system that has been developed to support analysis and improvement of teaching processes in e-learning environment University Rome Unitelma Sapienza. MILA offers variety real-time data visualizations provide statistics, trends, insight information both on Moodle-based virtual as whole, each course included it. In addition standard logs, is able analyze tracking generated by...