News Article

Brookhaven National Laboratory's synchrotron radiation facility, NSLS-II, is well known for its long-standing use of AI and machine learning for automated data analysis in beamlines.Automated anomaly detection, for example, is one of the more basic but useful applications of AI/machine learning.In a 24/7 research facility, when unmanned automated measurements are performed, the system detects when something unexpected happens and the data are abnormal or erroneous.To do this, an AI agent based on a supervised machine learning model was trained with a dataset of several hundred to several thousand examples of various abnormal or erroneous data that cannot be accepted in practice, and examples of normal data that can be accepted.Chatbot-based user support is also one of the introductory and very helpful applications.New users visiting the synchrotron beamlines can ask the chatbot questions about various problems they encounter (which may seem trivial to experienced users), and the chatbot responds immediately.Similar help functions are now being used in a wide range of areas, not only in research institutions, but also in society at large.