Gao et al. propose FedEmbryo, a federated task-adaptive learning (FTAL) approach that dynamically adjusts task and client contributions based on learning feedback, enabling privacy-preserving, decentralized training across multiple clinical sites to improve IVF outcomes. FedEmbryo demonstrates superior performance in both morphological evaluation and live-birth outcome prediction, thereby enhancing the accuracy of high-quality embryo selection.
- Tianrun Gao
- Yuning Yang
- Guangyu Wang
