← Back to all news
arXiv Computation and Language

Structuring occupational accident narratives for cross-sector safety analysis: Transferability of accident-process role classification

The study investigates whether a model trained on construction‑sector occupational accident narratives can accurately classify accident‑process roles in other sectors and reporting environments. Using 42,244 factual units from 6,040 construction narratives, the authors compared TF‑IDF, frozen pretrained representations, and task‑adapted pretrained models, achieving up to 85.7% balanced accuracy without retraining. The models performed consistently across metallurgy, chemistry‑plastics, and an independent company corpus, though performance varied more on the latter due to differing reporting practices.

Related stories

arXiv Computation and Language
Sep 21

Cross-sector generalization of accident-process role classification in occupational accident narratives

The study evaluates how well accident‑process role classifiers trained on construction‑sector French occupational accident narratives generalize to other sectors. An expert‑annotated corpus classifies factual units into four roles—work situation, unfavourable condition, accident event, and consequence—and the classifiers are tested on unseen metallurgy, chemistry‑plastics, and company corpora without retraining. Task‑specific adaptation consistently outperforms frozen representations, achieving balanced accuracies of 85.6%–85.8% across the target domains.

By Aho Yapi, Pierre Latouche, Arnaud Guillin, Yan Bailly
fine-tuning
More like this →
arXiv Computation and Language
Aug 25

ConstructCIE: A Dataset for Extracting Causal Information from Construction Accident Narratives

ConstructCIE is a manually annotated dataset designed for extracting causal information from OSHA construction accident reports. It employs a hierarchical schema that categorizes accident types, causal factors, sub‑causal factors, and the supporting evidence spans. Experiments with supervised sequence taggers and instruction‑tuned large language models show strong performance on accident‑type prediction and broad causal recovery, yet precise span‑level extraction remains challenging, highlighting the need for better domain grounding and evidence extraction.

By Hung Nguyen, Jaehoon Lee, Namgyun Kim, Kuan-Hao Huang
llms
More like this →
arXiv Machine Learning
Sep 16

Crash Narrative-Guided Countermeasure Recommendation Using Large Language Models: A Retrieval-Augmented Generation Framework for Intersection Safety

arXiv:2609.15997v1 Announce Type: cross Abstract: Improving safety at intersections requires identifying crash mechanisms and recommending appropriate countermeasures. However, this process tradition...

By Abu Saif Md Nasim Uddin, Mohamed Abdel-Aty, Zubayer Islam, Parvez Anowar, Chenzhu Wang
llmsragsafety
More like this →
arXiv AI
Sep 23

Toward Auditable and Calibrated AI for Dementia-Related Crash Severity Prediction: A Selective Deferral Framework to Support Human Review

The paper presents a decision‑aware framework for predicting dementia‑related crash severity that emphasizes auditability and selective deferral. Using 4,781 Texas crash records, the authors evaluate several models—including structured, narrative, fusion, calibrated fusion, BERT‑family, and local large‑language‑model baselines—under a stratified 70/15/15 split. The leakage‑controlled Gemma model achieves the highest macro‑F1 of 0.545, while a calibrated fusion model reaches 0.522 macro‑F1 with an expected calibration error of 0.033; selective deferral further improves performance, raising macro‑F1 to 0.573 at 70% coverage and reducing severity cost to 0.577.

By Gaurab Chhetri, Anika Baitullah, Subasish Das
safety
More like this →
arXiv Computation and Language
Sep 18

SAFARI: An Industrial Benchmark for LLM-Assisted Hazard Analysis and Risk Assessment

SAFARI is the first industrial benchmark for evaluating large language models (LLMs) in automotive hazard analysis and risk assessment (HARA) under ISO 26262. It comprises 3,000 de‑identified HARA cases and tests two tasks: open‑ended hazard generation and standards‑grounded risk classification, using a novel reference‑anchored LLM‑as‑a‑judge protocol. Experiments with nine state‑of‑the‑art LLMs show that while hazard narratives are often plausible, risk classification remains weak (best ASIL macro‑F1 = 0.261), with errors mainly due to missing scenario context and misjudged controllability. "whyItMatters":"The benchmark highlights the current limitations of LLMs in safety‑critical engineering workflows, guiding future research and expert oversight in automotive safety analysis."

By Chenxi Wu, Zimu Wang, Haiyang Zhang, Wei Wang, Zhijie Xu
llmsbenchmarkssafety
More like this →
arXiv AI
Jul 7

Shortcut Learning in Legal Judgment Prediction: Empirical Evidence from the UK Employment Tribunal

arXiv:2607. 04261v1 Announce Type: new Abstract: Current Legal Judgment Prediction (LJP) is constrained by its reliance on post-hoc judicial materials, increasing the likelihood that models perform retrospective classification rather than true forecasting.

By Joe Watson, Joana Ribeiro de Faria, Marcus Tomalin, M{\aa}ns Magnusson, Huiyuan Xie, Hao Tian Yeung, Felix Steffek
llms
More like this →