Nano Dimension Ltd. announced the filing of a U.S. patent application titled ?Large Language Models for Efficient Anomaly Detection in Log Files of Industrial Machines? (the ?Log Analysis Patent? or ?Patent?), which is targeted for real-time data analysis and scalable deployment across the Company?s own systems and industrial solutions provided to outside customers.

The Log Analysis Patent addresses one of the core challenges for automated anomaly detection. While machine logs are usually a valuable source of information for industrial systems, they are increasingly difficult and expensive to analyze as the underlying systems have grown in complexity and the volume of log data they contain has multiplied. Furthermore, logs are typically analyzed after events have happened and not in real-time, thereby missing the opportunity to apply corrective actions.

To overcome these problems, Nano Dimension has extended its existing AI patents with a Large Language Model (?LLM?) that can operate independently of engineering labels. With this, the technology exploits the existing sentiment that is expressed in the machine logs. This enables a fully automated process of AI-powered prediction of manufacturing anomalies before they occur, based either solely on logs or in combination with other machine data and being efficient enough to process billions of log lines.

The Log Analysis Patent follows a related patent that was filed and announced in September 2023. Both patents are equally valuable for the Company?s development of Nano Dimension?s products, which are being designed with DeepCube Group?s deep learning-based AI wherever possible, as well as for third party customers and partners, who are increasingly looking to the Company to leverage the same technology in their own industrial processes and systems.