The project COLABORATIVF aims to advance assisted reproductive technologies by developing a collaborative artificial intelligence platform to improve the success of embryo transfer in in vitro fertilization (IVF). The core technological objective is to design and validate AI models capable of analyzing large-scale, multimodal clinical data to identify reliable indicators of endometrial receptivity, a key yet poorly understood factor in IVF outcomes. The project builds on Seedchrony, an innovative medical device that measures intrauterine oxygen as a non-invasive biomarker, and integrates this signal with heterogeneous data sources such as clinical histories, embryology lab data, ultrasound images, and procedural parameters.

A major AI challenge addressed is data fragmentation and privacy in healthcare. To overcome this, the project adopts federated and distributed learning approaches, enabling multiple clinics and hospitals to collaboratively train models without sharing raw patient data. Additional challenges include managing high variability in clinical procedures, mitigating bias due to uneven data quality across centers, and ensuring model interpretability and regulatory compliance for medical AI. The platform also explores the use of large language models to extract structured information from unstandardized medical reports while preserving confidentiality. Ultimately, COLABORATIVF seeks to deliver scalable, ethical, and clinically robust AI solutions that enable personalized IVF treatments and shift the success metric from pregnancy to the birth of healthy children.