Vocab-Expander: A System for Creating Domain-Specific Vocabularies Based on Word Embeddings

Proceedings of Recent Advances in Natural Language Processing, pages 331–335 Varna, Sep 4–6, 2023 https://doi.org/10.26615/978-954-452-092-2_037 331 VOCAB-EXPANDER: A System for Creating Domain-Specific Vocabularies Based on Word Embeddings Michael F¨arber Karlsruhe Institute of Technology (KIT) Institute AIFB Karlsruhe, Germany michael.faerber@kit.edu Nicholas Popovic Karlsruhe Institute of Technology (KIT) Institute AIFB Karlsruhe, Germany popovic@kit.edu Abstract In this paper, we propose VOCAB-EXPANDER at https://vocab-expander.com, an on- line tool that enables end-users (e.g., technol- ogy scouts) to create and expand a vocabulary of their domain of interest. It utilizes an ensem- ble of state-of-the-art word embedding tech- niques based on web text and ConceptNet, a common-sense knowledge base, to suggest re- lated terms for already given terms. The system has an easy-to-use interface that allows users to quickly confirm or reject term suggestions. VOCAB-EXPANDER offers a variety of poten- tial use cases, such as improving concept-based information retrieval in technology and innova- tion management, enhancing communication and collaboration within organizations or inter- disciplinary projects, and creating vocabularies for specific courses in education. 1 Introduction Motivation. In many scenarios, it is necessary to create an ontology or other formal model of a domain of interest from scratch. For instance, in the field of technology and innovation manage- ment, technology scouts and other end-users with- out technical skills often use a list of terms for continuously retrieving and scanning texts from different media sources (e.g., news articles, social media, publications, patents) in order to become aware of novel relevant technologies and to create and populate profiles of technologies and actors within a particular domain, such as Smart Cities. However, coming up with such a vocabulary is typ- ically highly time-consuming and costly due to the domain-specificity (i.e., non-experts have no start- ing point what to add), the complexity of correctly defining the scope of the domain (e.g., Smart Cities can range from Smart Home to energy efficiency to security), the ambiguity of natural language (i.e., the meaning of terms may vary depending on the context in which they are used), and the emergence of new terms over time. Current Situation. So far, domain experts (e.g., technology scouts in technology and innovation management) still rely heavily on domain expert knowledge (de Weck, 2022). Several tools for mod- eling a domain of interest exist, including Proteg´e (Musen, 2015) and D-Terminer (Rigouts Terryn et al., 2022). However, these tools are often con- sidered as “too heavy” for creating only a domain- specific vocabulary instead of an ontology with a specific data model and standardizations (e.g., W3C RDF, OWL). Furthermore, these tools are typically designed to support the modeling process (e.g., based on an existing text corpus), but do not suggest directly related terms for given terms. Contributions. In this paper, we propose the system VOCAB-EXPANDER, available online at https://vocab-expander.com, that enables end- users without technical skills to create and expand a vocabulary of their domain of interest. The sys- tem utilizes an ensemble of state-of-the-art word embedding techniques to suggest related terms for already given terms. In addition to word …

Content truncated to 500 words