PHAT-JeT combines geometric message-passing with hierarchical patch attention to reach state-of-the-art accuracy and background rejection among resource-constrained jet tagging models on four benchmarks.
Fast inference of deep neural networks in FPGAs for particle physics
8 Pith papers cite this work, alongside 404 external citations. Polarity classification is still indexing.
8
Pith papers citing it
404
external citations · OpenAlex
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Differential halo zonotopes enable static verification of global robustness in DNNs by jointly propagating pairs of perturbed inputs while bounding divergence, with a relaxed confidence-based variant.
FPGA implementations for full matrix-element workflow on e+e- to mu+mu- and color-algebra kernels on gg to ttbar+X achieve speedups and energy gains over CPU/GPU while preserving numerical accuracy.
SAL-T enhances the linformer with spatially aware kinematic partitioning and convolutions to match full-attention transformer performance on jet tagging while keeping linear complexity and lower latency.
SNAC-Pack uses learned FPGA resource surrogates inside evolutionary neural architecture search to produce smaller, faster FPGA models for jet classification and qubit readout than BOP-based codesign.
WaveDriver is a laser guide star AO concept whose initial simulations indicate it may be required to meet HWO primary mirror segment stability and low-order wavefront stability requirements.
Reviews precision timing integration in LHC upgrades and discusses a possible shift to triggerless detectors enabled by timing and networking, with reflections on physics benefits.