Accelerating event reconstruction in neutrino telescopes using sparse convolutional neural networks
by
Tsung-Dao Lee Institute/N6F-N600 - Lecture Room
Tsung-Dao Lee Institute
Abstract:
Convolutional neural networks (CNNs) have seen extensive applications in scientific data analysis, including in neutrino telescopes. However, the data from these experiments present numerous challenges to CNNs, such as non-regular geometry, sparsity, and high dimensionality. In this talk, I will present sparse submanifold convolutions (SSCNNs) as a solution to these issues and show that the SSCNN event reconstruction performance is comparable to or better than traditional and machine learning algorithms. Additionally, our SSCNN runs approximately 16 times faster than a traditional CNN on a GPU. Finally, I will discuss our current efforts to implement this type of network into the IceCube Neutrino Observatory.
Brief Biography:
Felix Yu is currently a first-year physics PhD student at Harvard University, working within the IceCube Neutrino Observatory collaboration. Previously, he was an undergraduate at Tufts University, where he worked on deep learning techniques for the MicroBooNE neutrino experiment. His current research interests are focused on deep learning and its applications to high-energy physics. He has been actively involved in developing neural networks for data analysis in neutrino experiments, with a specific focus on making these methods as efficient and fast as possible.
Online Meeting Room: https://meeting.tencent.com/dm/aXzYIMqsrlVV
ID: 594-731-646 Password:123456
