Fast Jet Tagging with MLP-Mixers on FPGAs
This addresses the problem of efficient real-time data processing for particle physics experiments, representing a strong domain-specific advancement.
The paper tackled real-time jet tagging for particle colliders by using MLP-Mixer models on FPGAs, achieving state-of-the-art accuracy while reducing hardware resource usage by up to 97%, doubling throughput, and halving latency.
We explore the innovative use of MLP-Mixer models for real-time jet tagging and establish their feasibility on resource-constrained hardware like FPGAs. MLP-Mixers excel in processing sequences of jet constituents, achieving state-of-the-art performance on datasets mimicking Large Hadron Collider conditions. By using advanced optimization techniques such as High-Granularity Quantization and Distributed Arithmetic, we achieve unprecedented efficiency. These models match or surpass the accuracy of previous architectures, reduce hardware resource usage by up to 97%, double the throughput, and half the latency. Additionally, non-permutation-invariant architectures enable smart feature prioritization and efficient FPGA deployment, setting a new benchmark for machine learning in real-time data processing at particle colliders.