Papers by Junhyeong Park
Are Any-to-Any Models More Consistent Across Modality Transfers Than Specialists? (2025.acl-long)
Copied to clipboard
| Challenge: | equivariance evaluations uncover weak but observable consistency through structured analyses of the intermediate latent space enabled by multiple editing operations. |
| Approach: | They use a dataset of 1,000 images paired with captions, editing instructions, and Q&A pairs to evaluate cross-modal transfers rigorously. |
| Outcome: | The proposed models do not consistently demonstrate greater cross-modal consistency than specialized models in pointwise evaluations such as cyclic consistency. |
IntelliCAT: Intelligent Machine Translation Post-Editing with Quality Estimation and Translation Suggestion (2021.acl-demo)
Copied to clipboard
| Challenge: | Existing computer-aided translation tools require the translator to edit incorrect parts of a document, while ITP tools require fewer edits. |
| Approach: | They propose an interactive translation interface with neural models that streamline the post-editing process on machine translation output. |
| Outcome: | The proposed interface can significantly improve translation quality and a user study shows that it speeds up the post-editing process by 52.9% compared to translating from scratch. |