AI Medical Device Quality Standards: Terminology Overview

This draft standard outlines key AI, data management, and machine learning terms for medical device quality evaluation. Key points include:

Public Comment Deadline: 13-May-2025

Lead Organizations: China State Drug Administration, China Institute for Food and Drug Control

Focus: Aligns international and domestic terminology for AI device safety and performance evaluations.

Key Terms:

1. Deep Learning: A neural network approach with multiple hidden layers for input-output mapping.
2. Supervised Learning: Feedback-based learning for correctness verification.
3. Unsupervised Learning: Categorization without external feedback.
4. Reinforcement Learning: Reward-based learning for action maximization.
5. Edge Cloud Service: Cloud computing via edge computing facilities.
6. Data Acquisition: The process of generating and storing data digitally.
7. Data Masking/Anonymization: Protecting personal information by making it unidentifiable.
8. Artificial Neural Network (ANN): Processing elements linked by weighted connections.
9. Data Quality: Extent to which data meets requirements, like completeness and bias.
10. Training/Validation/Testing Data: Datasets for AI algorithm development and evaluation.

Region: Asia

Source: http://app.nifdc.org.cn/biaogzx/dataGkZqyj.do?formAction=viewZqyj&id=2c9048d8956b6a160195825f85435c51