Anomaly Finding in the LZ Dark Matter Experiment

The LUX-ZEPLIN (LZ) experiment, a 10-ton dual-phase xenon dark matter experiment at the SURF laboratory in South Dakota, is scheduled to start data collection in 2021. For a rare-event detector such as LZ, a general-purpose anomaly finder operating in a high-dimensional feature space can help in identifying and explaining the source of unexpected events stemming from rare backgrounds, rare signals, misclassifications, and abnormalities in detector function. If used with existing data selection algorithms, an anomaly finder can also increase the efficiency of identifying known backgrounds with unusual topologies. In this presentation, we will discuss the utility of anomaly finding and clustering to quickly identify outliers and their origins in simulated data, including those from rare background processes and data processing irregularities. We also discuss how to use these methods on experimental data when LZ comes online, with potential applications to data quality and data selection for specific analyses.