EUVIMEDEUVIMED entdecken
Crossref

Continual Learning for Resilient Multimodal Misinformation Detection Across Sequential Crisis Events

Bilal Ahmed Lodhi, Zeeshan Tariq

Proceedings of the AAAI Symposium Series · 23.06.2026 · 9 (1) · 91-97

DOI: 10.1609/aaaiss.v9i1.42910 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Symposium Series, online erschienen am 23.06.2026. Verfasst von Bilal Ahmed Lodhi und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Misinformation evolves rapidly during crises such as pandemics, political events and natural disasters, challenging the reliability of static detection systems. Although recent multimodal deep learning approaches have integrated text and imagery to improve misinformation detection, they have assumed stationary data distributions and overlooked the sequential domain shifts that are encountered in practice. In this study, we introduce a continual multimodal misinformation benchmark that evaluates resilience across three crisis domains: COVID-19 health misinformation, socio-political misinformation, and disaster-related credibility. We evaluated fine-tuning and replay-based continual learning strategies using a CLIP-based architecture to evaluate their performance. Our results reveal severe catastrophic forgetting under naive fine-tuning, with forgetting exceeding 12\%. In contrast, experience replay nearly eliminates forgetting, improves the average task accuracy by over 12 pp, and stabilizes predictive calibration. Our findings establish continual learning as a critical component for building resilient multimodal misinformation detection systems in dynamic real-world environments.

Zitieren

APA

Lodhi, B. A., & Tariq, Z. (2026). Continual Learning for Resilient Multimodal Misinformation Detection Across Sequential Crisis Events. Proceedings of the AAAI Symposium Series, 9(1), 91-97. https://doi.org/10.1609/aaaiss.v9i1.42910

Vancouver

Lodhi BA, Tariq Z. Continual Learning for Resilient Multimodal Misinformation Detection Across Sequential Crisis Events. Proceedings of the AAAI Symposium Series. 2026;9(1):91-97. doi:10.1609/aaaiss.v9i1.42910

BibTeX
@article{Lodhi2026,
  author = {Lodhi, Bilal Ahmed and Tariq, Zeeshan},
  title = {Continual Learning for Resilient Multimodal Misinformation Detection Across Sequential Crisis Events},
  journal = {Proceedings of the AAAI Symposium Series},
  year = {2026},
  volume = {9},
  number = {1},
  pages = {91--97},
  doi = {10.1609/aaaiss.v9i1.42910}
}
Kostenlos starten →