Reproduce “Lightweight Deep Learning Framework for Accurate Particle Flow Energy Reconstruction” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yu Wang, Yangguang Zhang, Shengxiang Lin, Xingyi Zhang, Han Zhang

The authors train CNNs to reconstruct particle-detector energy maps from sparse tracker and calorimeter images. They combine weighted mean-square error with a structural similarity loss and test channel attention and 3D convolutions. On simulated interpolation, a 92,465-parameter SEBlockCNN reaches WMSE 0.03093 versus 0.03187 for a 6,052,161-parameter CNN, but has worse jet-energy bias on held-out simulated jets.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.