Diagnostic Loss: Asymmetric Supervision for Radiology Report Generation
Automated radiology report generation requires models to do more than produce fluent clinical narratives: they must explicitly state supported diagnoses while avoiding unsupported ones. However, existing medical vision-language models (MedVLMs) are typically trained with token-level maximum likelihood objectives, which...
Ignacy Kolton, Wen-Xuan Li, Zong-Wei Zhou et al.
· 0 citations