Cargando…
The impact of inconsistent human annotations on AI driven clinical decision making
In supervised learning model development, domain experts are often used to provide the class labels (annotations). Annotation inconsistencies commonly occur when even highly experienced clinical experts annotate the same phenomenon (e.g., medical image, diagnostics, or prognostic status), due to inh...
| Autores principales: | , , , |
|---|---|
| Formato: | Online Artículo Texto |
| Lenguaje: | English |
| Publicado: |
Nature Publishing Group UK
2023
|
| Materias: | |
| Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9944930/ https://www.ncbi.nlm.nih.gov/pubmed/36810915 http://dx.doi.org/10.1038/s41746-023-00773-3 |