AIBIZ (AIBIZ, CEO Ha Seung-jae), a specialist in industrial AI (artificial intelligence), announced on the 11th that it has registered three patents for AI technologies that detect abnormal operation of process equipment. The newly registered patents are (1) 'a training method of a deep learning model for detecting abnormal operation of process equipment,' (2) 'a method of detecting abnormal operation of process equipment through analysis of a dynamic graph structure that changes over time,' and (3) 'a method of setting a reference threshold for detecting abnormal operation of process equipment.' According to the company, these technologies cover the entire process of learning, analyzing, and judging process data. It builds an anomaly detection model with deep-learning-based training technology, and detects anomaly signs that reflect changes in relationships between pieces of equipment and data flows through dynamic graph structure analysis. It is also designed to set reference thresholds suited to process characteristics so that AI diagnostic accuracy can be maintained stably. AIBIZ plans to strengthen the technological competitiveness of its AI-based manufacturing process diagnostics platform 'DutchBoy' through these patent registrations. The platform detects anomaly signals generated during manufacturing processes in real time and analyzes their causes, supporting improvements in productivity and yield. The company explained that it can be applied to a variety of process equipment environments, including high-precision manufacturing industries such as semiconductors, secondary batteries, and displays. A company official said, "We have secured the core elements of the AI technology needed for detecting process equipment anomalies," and "We will advance it into process diagnostics AI technology applicable across manufacturing."