Machine Learning in the LUX-ZEPLIN (LZ) Dark Matter Experiment

Over the past two decades, liquid xenon time projection chambers (TPCs) have emerged as the leading technology in direct dark matter detection, achieving unprecedented sensitivity to weakly interacting massive particles (WIMPs) and other particle-like candidates. The LUX-ZEPLIN (LZ) experiment, operating a dual-phase xenon TPC with a 7-tonne active mass, stands at the forefront of this effort. Machine learning methods offer powerful tools to enhance the experiment's sensitivity, given the complexity (more than 1,000 data-acquisition channels) and magnitude of the data collected (on the order of a petabyte annually). These applications span event reconstruction, background rejection, anomaly detection, and the use of large language models for knowledge retrieval. In this talk, I will present an overview of ML initiatives within LZ, outlining their scientific motivation, implementation, and impact.