@inproceedings{zhang-oussalah-2026-causal,
    title = "Causal Evidence Extraction and Triangulation in Crisis Reports using Large Language Models: A {R}elief{W}eb-based Study",
    author = "Zhang, Yuanjun  and
      Oussalah, Mourad",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1626/",
    pages = "32478--32491",
    ISBN = "979-8-89176-395-1",
    abstract = "Humanitarian reports are long, noisy, and multi-topic, making it difficult to consolidate decision-relevant causal evidence. We present a ReliefWeb study (2000{--}2024) and a two-stage Large Language Model (LLM) pipeline that extracts structured intervention-outcome records with direction and strength attributes. Query-conditioned extraction restricts output to a specified intervention class, reducing retrieval-induced over-extraction, while snippet grounding links each relation to supporting text for auditability and classification. In an expert-annotated dataset of 100 reports, the best closed-source LLM achieved a weighted F1 score of 90.73{\%} with strong cost-efficiency, while Llama-3.1-8B with supervised fine-tuning reached 94.15{\%} weighted F1 score. We further propose context-preserving triangulation that aggregates strength-weighted evidence within disaster$\times$source cells, applies Laplace smoothing and equally weights cells to quantify cross-context convergence via a Level-of-Evidence score. Applied to cash assistance, food-related outcomes show strong positive convergence (LoE=0.865) and stable long-horizon trajectories."
}