Ultrafast jet classification on FPGAs for the HL-LHC
This addresses the need for real-time data processing in high-energy physics experiments, but it is incremental as it applies existing methods to a specific hardware and data context.
The paper tackled the problem of performing ultrafast jet classification for the HL-LHC by optimizing machine learning models for deployment on FPGAs, achieving O(100) ns inference times with low resource costs.
Three machine learning models are used to perform jet origin classification. These models are optimized for deployment on a field-programmable gate array device. In this context, we demonstrate how latency and resource consumption scale with the input size and choice of algorithm. Moreover, the models proposed here are designed to work on the type of data and under the foreseen conditions at the CERN LHC during its high-luminosity phase. Through quantization-aware training and efficient synthetization for a specific field programmable gate array, we show that $O(100)$ ns inference of complex architectures such as Deep Sets and Interaction Networks is feasible at a relatively low computational resource cost.