Clinical implications of algorithmic interpretations of artificial intelligence in human embryo ploidy prediction
10.3760/cma.j.cn101441-20241105-00406
- VernacularTitle:以人工智能预测人类胚胎整倍性为例讨论算法解释性的意义
- Author:
Xiaodong ZHANG
1
;
Hao WANG
;
Xiaoni GUO
;
Wei HAN
;
Shubiao HAN
;
Guoning HUANG
Author Information
1. 重庆市妇幼保健院 重庆医科大学附属妇女儿童医院生殖医学中心，重庆 400010
- Publication Type:Journal Article
- Keywords:
Artificial intelligence;
Embryo;
Machine learning;
Deep learning;
Black-box;
Ploidy
- From:
Chinese Journal of Reproduction and Contraception
2025;45(1):31-38
- CountryChina
- Language:Chinese
-
Abstract:
Artificial intelligence (AI) technology is expected to assist physicians in improving the accuracy and efficiency of embryo assessment. However, embryo development is a continuous and dynamic process, when is meaningful or the whole development process need to be considered? Some research teams use static image analysis, which loses much important information, and others utilize algorithm-driven applications of "black-box" models to analyse embryo videos, which have limited their interpretability or explainability. Machine learning or deep learning is prone to abused due to its inherent complexity, and in order to apply AI more accurately, this paper discusses the clinical implications of algorithmic interpretations of AI in human embryo ploidy prediction.
