Towards tensor network models for low-latency jet tagging on FPGAs

Abstract We present a systematic study of tensor network (TN) models—matrix product states and tree TNs—for real-time jet tagging in high-energy physics, with a focus on low-latency deployment on field programmable gate arrays (FPGA). Motivated by the strict requirements of the high-luminosity large hadron collider Level-1 trigger system, we explore TNs as compact and interpretable alternatives to deep neural networks. Using low-level jet constituent features, our models achieve competitive performance compared to state-of-the-art deep learning classifiers. We investigate post-training quantization to enable hardware-efficient implementations without degrading classification performance or latency. Selected models are synthesized for FPGA deployment and implemented to obtain resource usage, latency, and memory occupancy evaluation. The resulting measurements and timing estimates indicate sub-microsecond inference latency, supporting the feasibility of TN-based models for online deployment in real-time trigger systems. Overall, this study highlights the potential of TN-based models for fast and resource-efficient inference in low-latency environments.