The author presents a concept-based interpretable model for diagnosing choroid neoplasias, emphasizing the importance of interpretability in medical AI.