Deep learning model segments and classifies particle decays efficiently
ResoSeg: Resonance Tagger using Transformer and Segment Model
Artificial Intelligence
Summary
The paper addresses a challenge in particle physics where scientists study particles that decay into many different things. The authors present ResoSeg, a deep learning model that can look at all the pieces from a particle collision and both separate each particle's tracks and recognize which particle resonance they came from. This approach works faster and more accurately than previous methods that needed multiple steps and specific setups for each type of decay. The model also adapts well to different energy conditions and particle types, making it broadly useful in particle experiments.
What this means in practice
- •For particle physics experimental teams: Process collision data to identify and reconstruct particles in one step, improving efficiency across resonance decay channels.
- •For medical imaging researchers: Use the joint segmentation and classification approach to improve identification of structures in complex imaging data.
Authors
Chunkai Li, Junhao Yin, Ke Li, Jingde Chen
Abstract
Deep learning has been widely applied across many areas of experimental high-energy physics, yet existing models address only event-level classification or object tagging and therefore still require reconstruction algorithms tailored to each decay channel. We present the first application of segmentation to resonance tagging at BESIII and introduce ResoSeg, a deep learning model that jointly performs particle-level segmentation and event-level classification, enabling a one-pass analysis of resonance to anything decays while precisely reconstructing the relevant resonance properties. We demonstrate the reconstruction of $η_c$ with $e^+e^-\toπ^+π^-h_c$, $h_c\toγη_c$, $η_c\to\text{anything}$. The model is trained on BESIII-$η_c$ dataset, which is constructed with per-track true labels obtained via a Truth-Matching Algorithm. Experimental results show that the average combined efficiency of ResoSeg is more than double that of the conventional 16-channel approach across energy points from 4.19 to 4.60\,GeV. The model generalizes to unseen energy points, adapts to other $η_c$ production modes through transfer learning, and remains robust against variations in the $η_c$ mass, width, and branching fractions, providing a general, resonance-aware model applicable beyond $η_c$ and BESIII. The source code is available at https://github.com/oashen/ResoSeg.