The authors created an expert-annotated French corpus whose factual units were assigned one of four roles: work situation (A0), explicitly reported unfavourable condition (A1), accident event or deviation (B), and reported consequence (C).
Accident classifiers transfer across industries after task-specific adaptation
Models trained only on French construction reports classified accident-process roles in three unseen industrial corpora with roughly 86% balanced accuracy.
Academic
Aho Yapi · Pierre Latouche · Arnaud Guillin · Yan Bailly
Laboratoire de Mathématiques Blaise Pascal (UMR 6620 CNRS) · Université Clermont Auvergne · LYF SAS · Cikaba · Institut Universitaire de France (IUF)
Research Digest··2 min read
Yapi and colleagues tested whether classifiers trained on occupational accident narratives from one sector could work in others without retraining.
Why this paper
From Laboratoire de Mathématiques Blaise Pascal (UMR 6620 CNRS) and 4 others
In one line
A model trained on construction accident narratives identifies accident-process roles across sectors and reporting environments.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§