State-of-the-art models in lexical semantic change detection
Event date:  January 18, 2021, 4:00 PM  to 6:00 PM

Abstract

Lexical semantic change detection is a relatively young field in computational linguistics concerned with the automatic detection of meaning changes in words over time. I will first introduce low-dimensional semantic vector space models (VSMs) which dominate the task within the current evaluation framework. Then, I will distinguish between type- and token-based VSMs and discuss possible reasons for the apparent dominance of the simpler type-based models. I will show that token-based models can be tuned to reach similar performance as type-based models, while losing much of their expressiveness. Both model types then yield reasonable performance on current data sets. However, I will argue that this performance cannot be fully trusted, as these data sets are strongly biased towards polysemy. I will end by giving an outlook towards a formulation of semantic change detection as supervised learning task, which I consider to be the most promising direction of future research.


About the Speaker

I’m a last-year PhD student at the IMS (University of Stuttgart) working together with Sabine Schulte im Walde on automatic detection of lexical semantic change. I hold a Bachelor degree in Linguistics and English and a Master degree in Computational Linguistics. My interest is focused on the application of machine learning methods to solve problems involving the semantics of words.


URL: https://www.ims.uni-stuttgart.de/en/institute/team/Schlechtweg/


About the Series

The Lecture Series is organized every winter term and consist of talks with international speakers. The speakers are experts from various fields and professions. The talks are aimed at covering the whole spectrum of visual computing and at discussing the relevance of quantification. This years lecture series will be held entirely online.

This is an online event only.

Registration: If you would like to attend, please email Leonel Merino to register.