Speaker
Description
The COherent Neutrino–Nucleus Interaction Experiment (CONNIE) operates 30 meters from the Angra 2 reactor at the Almirante Álvaro Alberto Nuclear Power Plant in Brazil. CONNIE is the first reactor neutrino experiment to employ silicon Skipper-CCDs with the aim of detecting coherent elastic neutrino-nucleus scattering (CEvNS), as well as searching for new physics, and enabling real-time reactor monitoring. A significant challenge lies in reliably identifying the diverse event types within extensive data collected by the detector. To address this, we propose the use of classical machine learning algorithms and convolutional neural networks to categorize events, benchmarking these approaches. A labeled dataset of experimental events was developed using the Annotation Redundancy with Targeted Quality Assurance method, where multiple annotators independently label overlapping data subsets to identify and resolve inconsistencies, leading to a reliable and robust dataset that can be used for training. This work presents recent results from the classification models and shares progress in developing the annotated dataset.
| I read the instructions above | Yes |
|---|