The Past, Present, and Future of NLP from a Linguistic Perspective
Since the first modern Neural Networks were designed in the 1980s to, among other tasks, model English past tense inflection, Natural Language Processing and the Linguistic debate about the modelling of human language have mostly diverged. Nevertheless, they have interacted and shaped each other in important ways at several key moments in the past, and are interacting again in the present, in the discussion about whether state-of-the-art language models reach their performances the ‘right‘ way. This talk will explore the development of NLP from the perspective of the Linguistic debate of rule-l vs. usage-based theories, and highlight why taking a stand in this debate is important for finding solutions to the challenges facing state-of-the-art language models.
Leonie Weissweiler is a PhD student at the Center for Information and Language Processing at LMU Munich, supervised by Prof. Dr. Hinrich Schütze. She is interested in leveraging methods from Natural Language Processing to contribute to the empirical study of the emergent structure of Language, its evolution and processing in the brain. Her current research focuses on multilingual unsupervised Morphosyntax and probing language models for Construction Grammar.