EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Zero-Shot Social Media Crisis Classification: A Training-Free Multimodal Approach

Franziska Schwarz, Klaus Dieter Schwarz, Daniel Arias Aranda, Kendrick Bollens, Navaneeth Shivananjappa, Reiner Creutzburg, Vesna Dimitrova

Applied Sciences · 25.02.2026 · 16 (5) · 2192

DOI: 10.3390/app16052192 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 25.02.2026. Verfasst von Franziska Schwarz und 6 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union in the framework of 838 ERASMUS MUNDUS.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Rapid classification of social media content during humanitarian crises is essential for effective disaster relief; however, traditional approaches require extensive annotated training data, which are often unavailable during new disasters. This paper presents a training-free, multimodal classification framework that leverages zero-shot vision-language models to analyze disaster-related social media content without task-specific training. The framework employs a two-stage prompt-engineered pipeline using the locally deployable Mistral-Small-3.1-24B-Instruct model, performing binary informativeness detection followed by multiclass categorization into eight humanitarian categories through structured JSON output generation. Evaluation on the CrisisMMD dataset of 18,082 multimodal samples from seven natural disasters demonstrated binary F1 scores above 0.84 for both text and image informativeness detection and weighted F1 scores of 0.61 (text) and 0.72 (image) for humanitarian categorization. The framework generalizes consistently across all disaster types with minimal performance variance (standard deviation below 0.031) and operates entirely on local infrastructure without cloud dependencies, requiring only moderate GPU resources. By eliminating training data requirements, this approach enables immediate deployment during new disasters, demonstrating that zero-shot multimodal classification achieves practically relevant performance for real-time crisis response.

Förderung

  • European Union in the framework of 838 ERASMUS MUNDUS

Zitieren

APA

Schwarz, F., Schwarz, K. D., Aranda, D. A., Bollens, K., Shivananjappa, N., Creutzburg, R., & Dimitrova, V. (2026). Zero-Shot Social Media Crisis Classification: A Training-Free Multimodal Approach. Applied Sciences, 16(5), 2192. https://doi.org/10.3390/app16052192

Vancouver

Schwarz F, Schwarz KD, Aranda DA, Bollens K, Shivananjappa N, Creutzburg R, et al. Zero-Shot Social Media Crisis Classification: A Training-Free Multimodal Approach. Applied Sciences. 2026;16(5):2192. doi:10.3390/app16052192

BibTeX
@article{Schwarz2026,
  author = {Schwarz, Franziska and Schwarz, Klaus Dieter and Aranda, Daniel Arias and Bollens, Kendrick and Shivananjappa, Navaneeth and Creutzburg, Reiner and Dimitrova, Vesna},
  title = {Zero-Shot Social Media Crisis Classification: A Training-Free Multimodal Approach},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {5},
  pages = {2192},
  doi = {10.3390/app16052192}
}
Kostenlos starten →