On-chip probabilistic inference for charged-particle tracking at the sensor edge

TL;DR

This paper demonstrates neural network-based probabilistic inference for charged-particle tracking directly at the sensor edge, enabling efficient data processing under strict constraints in high-rate scientific instruments.

Contribution

It introduces embedded neural networks for real-time particle kinematic inference on silicon sensors, meeting tight precision, latency, and area requirements.

Findings

Neural networks can accurately regress hit positions and angles with calibrated uncertainties.

The approach satisfies constraints on numerical precision, latency, and silicon area.

It paves the way for probabilistic inference directly at the sensor edge in scientific detectors.

Abstract

Modern scientific instruments operate under increasingly extreme constraints on bandwidth, latency, and power. Inference at the sensor edge determines experimental data collection efficiency by deciding which information to save for further analysis. Particle tracking detectors at the Large Hadron Collider exemplify this challenge: pixelated silicon sensors generate rich spatiotemporal ionization patterns, yet most of this information is discarded due to data-rate limitations. Concurrently, advancements in co-design tools provide rapid turn-around for incorporating machine learning into application-specific integrated circuits, motivating designs for particle detectors with new integrated technologies. We demonstrate that neural networks embedded in the front-end electronics can infer charged-particle kinematic parameters from a single silicon layer. We regress hit positions and incident angles with calibrated uncertainties, while satisfying stringent constraints on numerical precision, latency, and silicon area. Our results establish a path toward probabilistic inference directly at the edge, opening new opportunities for intelligent sensing in high-rate scientific instruments.