From Invisible to Visible: (Anti-)Neutron Reconstruction with deep learning

by Dr Yunxuan Song（宋昀轩） (École Polytechnique Fédérale de Lausanne (EPFL))

Tuesday 30 Sept 2025, 12:00 → 13:30 Asia/Shanghai

5#/6th-603 - Meeting Room 603 (Science Building)

Abstract:

Long-lived neutral hadrons, such as (anti-)neutrons, serve as important probes in flavour physics. However, most flavour-physics experiments lack dedicated hadronic calorimeters, and neutral-hadron detection must instead rely on the electromagnetic calorimeter (EMC). Due to the EMC’s limited volume and dense material, hadronic showers are only partially contained, posing significant challenges for conventional reconstruction methods. In this talk, we present the Vision Calorimeter (ViC) and Language Calorimeter (LaC), two deep-learning frameworks inspired by recent advances in computer vision and natural language processing. By adopting an end-to-end, data-driven strategy, ViC and LaC achieve unified reconstruction of anti-neutrons, simultaneously performing particle identification, incident-position estimation on the EMC, and momentum-magnitude inference.

Biography:

Dr. Yunxuan Song is a scientist at the École Polytechnique Fédérale de Lausanne (EPFL) and a member of the LHCb and BESIII collaborations. His research focuses on flavor physics, CP violation, spectroscopy, and searches for new physics, with a strong emphasis on novel methodologies and advanced machine learning techniques in experimental particle physics. He received his Ph.D. in Physics from Peking University and was awarded the BESIII PhD Thesis Award in recognition of his outstanding contributions to the experiment.

Host: Dr. Xun Chen

Alternative online link：https://meeting.tencent.com/dm/iv1smHdmuBmZ

Id: 472421801   password: 123456