Papers by Brandon Lwowski
Measuring Geographic Performance Disparities of Offensive Language Classifiers (2022.coling-1)
Copied to clipboard
| Challenge: | Recent work shows that text classifiers are biased regarding different languages and dialects. |
| Approach: | They propose to use a dataset to examine whether language, dialect, and topical content vary across geographical regions to address these gaps. |
| Outcome: | The proposed dataset includes 14 thousand examples across 15 cities and shows that current models do not generalize across locations. |
An Empirical Study of the Downstream Reliability of Pre-Trained Word Embeddings (2020.coling-main)
Copied to clipboard
| Challenge: | Pre-trained word embeddings have been shown to improve the performance of neural networks across a wide variety of tasks. |
| Approach: | They propose two new metrics to understand the downstream reliability of word embeddings. |
| Outcome: | The proposed model can improve performance with slight changes to the training data, but it can also fail with multiple neural network architectures. |