Papers by Arezoo Hatefi
PromptStream: Self-Supervised News Story Discovery Using Topic-Aware Article Representations (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing methods for news story discovery relied on sparse document representations such as keywords and TF-IDF vectors. |
| Approach: | They propose a method that constructs article embeddings using cloze-style prompting and self-supervised contrastive learning techniques to tackle this task. |
| Outcome: | The proposed model is able to identify coherent news stories within a news stream and to monitor their progress. |