Accelerating Deep Neural Networks for Real-time Data Selection for High-resolution Imaging Particle Detectors

TL;DR

This paper develops optimized FPGA implementations of convolutional neural networks to enable real-time, power-efficient data selection in high-resolution particle detectors like DUNE LArTPC, improving data processing for neutrino experiments.

Contribution

It introduces a custom FPGA-based approach for accelerating neural network inference tailored to high-resolution particle detector data, addressing power and speed constraints.

Findings

Achieved real-time inference speeds on FPGA hardware.

Demonstrated power efficiency suitable for remote detector operation.

Validated performance on simulated detector data.

Abstract

This paper presents the custom implementation, optimization, and performance evaluation of convolutional neural networks on field programmable gate arrays, for the purposes of accelerating deep neural network inference on large, two-dimensional image inputs. The targeted application is that of data selection for high-resolution particle imaging detectors, and in particular liquid argon time projection chamber detectors, such as that employed by the future Deep Underground Neutrino Experiment. We motivate this particular application based on the excellent performance of deep neural networks on classifying simulated raw data from the DUNE LArTPC, combined with the need for power-efficient data processing in the case of remote, long-term, and limited-access operating detector conditions.