Predicting Scalar Inference and Alternativehood With Neural Sentence Encoders
Abstract/Contents
- Abstract
- Scalar inference is the interpretation of a less informative term to also mean the negation of a more informative alternative. The strength of these inferences is often sensitive to context and the presence of linguistic cues. Neural sentence encoders have been shown to successfully predict the strength of certain scalar inferences (in particular, from some to not all). In this work, we explore two topics: First, the extent of variability in scalar inference strength from or to not both, and whether this inference strength can be predicted by a neural sentence encoder; second, whether a BERT-based framing of lexical expectation can be used to model the availability of scalar alternatives to or. We find that a BERT-based LSTM predicts inference strength from or to not both with limited success (r = 0.45), due to the relative pragmatic complexity of or. We also find that a BERT- based language model successfully predicts the categorical availability of alternatives for or, indicating that BERT captures qualitative information about alternativehood, and that lexical expectation measures of alternativehood weakly predict inference strength.
Description
Type of resource | text |
---|---|
Date created | June 3, 2021 |
Creators/Contributors
Author | Li, Elissa |
---|
Subjects
Subject | neural network |
---|---|
Subject | BERT |
Subject | alternatives |
Subject | scalar inference |
Subject | pragmatics |
Subject | machine learning |
Subject | NLP |
Subject | Symbolic Systems |
Genre | Thesis |
Bibliographic information
Access conditions
- Use and reproduction
- User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Preferred citation
- Preferred Citation
- Li, Elissa. (2021). Predicting Scalar Inference and Alternativehood With Neural Sentence Encoders. Stanford Digital Repository. Available at: https://purl.stanford.edu/xm864rp2970
Collection
Master's Theses, Symbolic Systems Program, Stanford University
View other items in this collection in SearchWorksContact information
- Contact
- eyl2125@gmail.com
Also listed in
Loading usage metrics...