This is the project that I'm currently working on over summer 2023!
in the presence of a static external magnetic field, Axion Dark Matter behaves like an effective current density. The ABRACADABRA Experiment was built to detect axion-like particles through capturing this effective flux with superconducting quantum interference devices (SQUID) in a toroidal magnet.
The current data analysis procedure to identify those excess power caused by the axions involves Fourier transforming the data, and down-sampling multiple times, which causes a lot of information loss. I have been working on preparing time series data for training a machine learning model to identify interesting frequencies directly from the time series data. I investigated the different model options and created a pipeline for digitally injecting axion signals into time series data, which allows us to generate useful training data for the ML mode.
Still a work in progress...