When Less is More: Towards Lightweight and Distilled Graph Neural Networks for Efficient Particle Reconstruction

Lightweight, attention-enhanced graph neural networks built on node-centric GarNet layers for real-time particle reconstruction and identification in LHCb’s next-generation calorimeter. The architecture eliminates costly edge message passing and achieves up to 8× faster inference than message-passing GNNs while maintaining superior energy resolution.

Sep 8, 2025 — Sep 12, 2025