A Computational Exploration of the Creative Aspect of Language

Israel Ministry of Science & Technology research grant, Digital Humanities (320,000 NIS for 3 years)

Dr. Nurit Melnik, The Department of Language, Literature, and the Arts at the Open University.
Dr. Gabriel Stanovsky, The School of Computer Science & Engineering at The Hebrew University of Jerusalem.

The notion of creativity is often invoked in the linguistics literature. Almost every introductory textbook emphasizes the human ability to produce and understand sentences that have never been made or heard before. This, it is assumed, is one property that makes human language unique and sets it apart from nonhuman communication systems. Nevertheless the notion of creativity is difficult to pin down in general, and in the field of linguistics its definition largely depends on how the language system is conceptualized.

Recent advances made in the fields of natural language processing (NLP), machine learning, and deep learning have made it possible to computationally explore language use and gain new insights. Deep learning architectures, relying on word embeddings and contextual language models, are successfully used for downstream tasks such as sentiment analysis. In this interdisciplinary project we will employ state-of-the-art computational tools and methods to conduct a data-driven exploration of the creative aspects of language use in order to advance our understanding of the nature of language.