Artificial Intelligence (AI)	Computer programs designed to mimic human thinking and decision-making, such as recognizing patterns in medical images or predicting patient risks.
Machine Learning (ML)	A subset of AI where computers “learn” from large amounts of data (like past patient records) to make predictions or recommendations without being explicitly programmed for every scenario.
Convolutional Neural Network (CNN)	A type of computer model especially good at analyzing medical images—think of it as software that “looks” at ultrasound or embryo pictures and highlights important features.
Recurrent Neural Network (RNN)/Long Short-Term Memory (LSTM)	Computer models that work well with data that comes in a sequence over time (for example, heart-rate changes during labour). They help predict events (such as when labor might start).
Transformer Model (e.g., ResNet, U-Net)	Advanced computer systems that process information (like ultrasound frames) very quickly and can point out abnormalities—imagine a tool that instantly draws outlines around the fetus in an ultrasound image.
Area Under the Curve (AUC)	A score (from 0 to 1) that tells us how well a test or model can tell “sick” versus “healthy.” An AUC of 0.80 means the tool is correct 80% of the time at distinguishing disease.
Dice Coefficient (also called Dice Score)	A number (0 to 1) that measures how closely a computer’s outline of an organ (for example, the fetal head) matches an expert’s outline. A Dice score of 0.90 means 90% overlap—very close agreement.
Sensitivity	The ability of a test or model to correctly identify those who have the condition. If sensitivity is 93%, it catches 93 out of 100 true cases.
Mean Absolute Error (MAE)	The average amount by which a prediction (such as fetal age in weeks) differs from the true value. If MAE is less than 1 week, the prediction is usually within one week of the actual age.
Clinical Decision Support System (CDSS)	A software tool that provides healthcare workers with personalized recommendations—such as reminding a midwife when to screen for postpartum depression or alerting a nurse to possible neonatal sepsis.
Embryo Selection Platform (e.g., ERICA, STORK-A, KIDScore, iDAScore)	Computer tools used in fertility clinics to choose the healthiest embryo(s) by analyzing images and data, aiming for higher chances of successful pregnancy.
Clinical Practice Guidelines (e.g., TRIPOD-AI, CONSORT-AI)	Checklists and rules for how to report and evaluate AI tools so that doctors and nurses can trust they work safely and fairly.
Electronic Health Record (EHR)/Laboratory Information Management System (LIMS)	Digital systems that store patient information and lab data. When AI is “integrated” with EHRs/LIMS, it means these tools can automatically pull and analyze patient data without extra manual steps.
Federated Learning	A way for hospitals to train AI tools on their own patient data without sharing the actual data with each other—only the “lessons learned” go back to a central system. This protects privacy while improving model accuracy across different regions.