Sara Carvalho, Águeda School of Technology and Management, University of Aveiro / NOVA FCSH
Terminology and linguistic linked data in the biomedical domain: hype, hope and reality
Over the last few decades, healthcare has been impacted by a series of critical changes, mostly from a technological standpoint, which, in some cases, have led to hurdles concerning the dissemination of clinical data across multiple providers, institutions, countries, and stakeholders. With its interdisciplinary nature, Terminology can make a valuable contribution in this regard by supporting a more effective knowledge representation, organization, and sharing in this field, thereby enhancing specialized communication among these various stakeholders. More specifically, it is believed that the potential resulting from the combined efforts of a double-dimensional (conceptual and linguistic) framework of terminology work and the more recent Linguistic Linked Data paradigm can help address some of the current challenges, namely those inherent to semantic interoperability, while also paving the way for new opportunities in linking different resources and making them available to both human and machine end users.
Špela Vintar, University of Ljubljana
Frames, knowledge models and cognitive spaces: Do they overlap?
Frames have gained popularity in terminology because of their potential of reconciling the cognitive, textual and formal views of specialized knowledge. As a result, the number of terminological knowledge bases employing the frame-based approach is growing, with emerging synergies between semantic web technologies and terminological knowledge resources. With recent advances in Natural Language Processing, new methods of knowledge modelling and knowledge mining are being explored, hoping to facilitate the construction of linked specialized resources.
In the first part of the talk we present the construction of a new knowledge base for karstology which employs the frame-based approach, combines manual and NLP methods to identify elements of knowledge, and includes graphs, images and maps into the representation of concepts. In the second – more experimentally oriented – part, we concern ourselves with the question whether semantic relations as stipulated through frames, and semantic similarity as represented through nearest neighbors in word embeddings, coincide or overlap with cognitive spaces as attested through free word associations. We present the results of two experiments where the psycholinguistic method of free associations was used to collect responses from expert and lay participants. In the first experiment we analyse the reponses in search of implicit frames, and in the second we compare them to the semantically similar words computed through word embeddings. We conclude by interpreting the interesting findings from a multidisciplinary perspective, leaving plenty of room for discussion.
Friedemann Vogel, University of Siegen
How to make the law more understandable – for whom? Results from an empirical evaluation of the linguistic legal editorial in the German Federal Ministry of Justice
It is an old question both in law and linguistics: how can we improve a text of legal norms and provide a better basis for the practice of legal interpretation not only in law but also in non-legal contexts? The main problem with this question and the answers given so far is that there are mostly only hypothetical assumptions about the addressees and actual recipients of legal texts. The lecture will present some results of an empirical project evaluating linguistic legal editing in German legislation (2019-2021), which will provide insight into the opportunities and challenges of developing a more comprehensible and addressee-oriented law.