Papers by Sonja Horn
PAUSE: Positive and Annealed Unlabeled Sentence Embedding (2021.emnlp-main)
Copied to clipboard
Lele Cao, Emil Larsson, Vilhelm von Ehrenheim, Dhiana Deva Cavalcanti Rocha, Anna Martin, Sonja Horn
| Challenge: | Sentence embedding is a set of effective and versatile techniques for converting raw text into numerical vector representations. |
| Approach: | They propose a generic and end-to-end approach to embed sentences from a partially labeled dataset using supervised methods. |
| Outcome: | The proposed approach achieves state-of-the-art results using only a small fraction of labeled sentence pairs on various benchmark tasks. |