Pulse shape descrimination

For applications such as the characterization of nuclear physics interactions and the safeguarding of nuclear materials, the classification of radiation sources is of significant interest. We have developed both unsupervised and semi-supervised variational autoencoder (VAE) models for the classification of radiation sources. VAEs, distinct from traditional classifiers, enable non-linear dimensionality reduction and uncertainty quantification, as detailed in AbdulazizICASSP_2022.