Papers by Megan Ayers
Discovering influential text using convolutional neural networks (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for estimating the effects of text on human evaluation are limited to testing a small number of pre-specified text treatments. |
| Approach: | They propose a method for flexibly discovering clusters of similar text phrases that are predictive of human reactions to texts using convolutional neural networks. |
| Outcome: | The proposed method can detect and predict human reactions to texts under certain assumptions. |