Generalizing to Unseen Disaster Events: A Causal View
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918200503959552 |
|---|---|
| author | Seeberger, Philipp Freisinger, Steffen Bocklet, Tobias Riedhammer, Korbinian |
| author_facet | Seeberger, Philipp Freisinger, Steffen Bocklet, Tobias Riedhammer, Korbinian |
| contents | Due to the rapid growth of social media platforms, these tools have become essential for monitoring information during ongoing disaster events. However, extracting valuable insights requires real-time processing of vast amounts of data. A major challenge in existing systems is their exposure to event-related biases, which negatively affects their ability to generalize to emerging events. While recent advancements in debiasing and causal learning offer promising solutions, they remain underexplored in the disaster event domain. In this work, we approach bias mitigation through a causal lens and propose a method to reduce event- and domain-related biases, enhancing generalization to future events. Our approach outperforms multiple baselines by up to +1.9% F1 and significantly improves a PLM-based classifier across three disaster classification tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_10120 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Generalizing to Unseen Disaster Events: A Causal View Seeberger, Philipp Freisinger, Steffen Bocklet, Tobias Riedhammer, Korbinian Computation and Language Machine Learning Due to the rapid growth of social media platforms, these tools have become essential for monitoring information during ongoing disaster events. However, extracting valuable insights requires real-time processing of vast amounts of data. A major challenge in existing systems is their exposure to event-related biases, which negatively affects their ability to generalize to emerging events. While recent advancements in debiasing and causal learning offer promising solutions, they remain underexplored in the disaster event domain. In this work, we approach bias mitigation through a causal lens and propose a method to reduce event- and domain-related biases, enhancing generalization to future events. Our approach outperforms multiple baselines by up to +1.9% F1 and significantly improves a PLM-based classifier across three disaster classification tasks. |
| title | Generalizing to Unseen Disaster Events: A Causal View |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2511.10120 |