Speaker
Description
In this talk I will present updated results regarding the application of machine learning techniques for noise removal and physics-variable prediction on signals from spherical proportional counters (SPCs) with the NEWS-G experiment. In SPC detectors, a primary ionization, created by a particle interacting with the gas, drifts towards a central anode. When ions approach the anode, the electric field becomes strong enough to trigger secondary ionizations, resulting in an amplified detector signal. Evaluation of these techniques include tests on simulated pulses with added noise and quantifying the impact of noise-removal and single-output prediction on physics goals such as primary ion counting and energy resolution. Successful implementation of this technique will reduce errors on event measurements (energy, drift time, etc.) and lower the analysis threshold, thereby enabling the experiment to search for lower mass dark matter events.
| Keyword-1 | Dark Matter |
|---|---|
| Keyword-2 | Machine Learning |
