This paper describes the system Clustering on Manifolds of Contextualized Embeddings (CMCE) submitted to the SemEval-2020 Task 1 on Unsupervised Lexical Semantic Change Detection. Subtask 1 asks to identify whether or not a word gained/lost a sense across two time periods. Subtask 2 is about computing a ranking of words according to the amount of change their senses underwent. Our system uses contextualized word embeddings from MBERT, whose dimensionality we reduce with an autoencoder and the UMAP algorithm, to be able to use a wider array of clustering algorithms that can automatically determine the number of clusters. We use Hierarchical Density Based Clustering (HDBSCAN) and compare it to Gaussian Mixture Models (GMMs) and other clustering algorithms. Remarkably, with only 10 dimensional MBERT embeddings (reduced from the original size of 768), our submitted model performs best on subtask 1 for English and ranks third in subtask 2 for English. In addition to describing our system, we discuss our hyperparameter configurations and examine why our system lags behind for the other languages involved in the shared task (German, Swedish, Latin
Fiction reading is a popular leisure activity associated with a variety of pleasurable experiences, including suspense, narrative transportation, and—as indicated by recent empirical studies—also flow. In the context of fiction reading, flow—generally defined as a pleasurable state of mind experienced during an optimally stimulating activity—is specifically related to an optimal balance between text‐driven challenges and the reader’s capabilities in constructing a mental story model. The experimental study reported here focused on the psychophysiological underpinnings of flow in the reading context. Cardiovascular data were collected from 84 participants both during a relaxation baseline prior to reading and during reading. Participants were randomly assigned to read one of three versions of a chapter from Homer’s Odyssey. According to statistical readability indices, these versions were low, intermediate, or high in readability, and hence in cognitive challenge. Flow was measured immediately after reading with a self‐report scale that was tailored to assess reading‐specific flow experiences. Regression analyses revealed that cardiovascular activation patterns measured before reading that are reflective of parasympathetic dominance—that is, an inner state associated with relaxation and cognitive fluency—moderated flow experiences during reading. In line with the stipulations of flow theory in regard to matching challenge levels being the key determinant for flow, this pattern supported subsequent flow experiences only in response to text versions of high or intermediate, but not of low cognitive challenge. Differences in cardiac vagal tone during reading were, however, not sensitive to our experimental modifications and not predictive of flow experiences.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.