Papers by Benjamin Swanson
BinaryAlign: Word Alignment as Binary Sequence Labeling (2024.acl-long)
Copied to clipboard
| Challenge: | State-of-the-art word alignment training methods require a different class depending on the availability of gold data for a particular language pair. |
| Approach: | They propose a novel word alignment technique based on binary sequence labeling that outperforms existing approaches in both scenarios. |
| Outcome: | The proposed method outperforms existing models on non-English language pairs and performs stratified error analysis over alignment error type. |
Zero-shot Cross-Lingual Transfer for Synthetic Data Generation in Grammatical Error Detection (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for Grammatical Error Detection (GED) rely on human annotations, but these are unavailable in many low-resource languages. |
| Approach: | They propose a two-stage fine-tuning pipeline to train a GED model using synthetic errors from target languages and human-annotated GED corpora from source languages. |
| Outcome: | The proposed method outperforms current state-of-the-art annotation-free GED methods and produces errors that are more diverse and similar to human errors. |