Development of a deep learning based automated data analysis for step-filter x-ray spectrometers in support of high-repetition rate short-pulse laser-driven acceleration experiments.

R A Simpson,D Mariscal,G G Scott,E Grace,T Ma,G J Williams

doi:10.1063/5.0043835

Abstract

We present a deep learning based framework for real-time analysis of a differential filter based x-ray spectrometer that is common on short-pulse laser experiments. The analysis framework was trained with a large repository of synthetic data to retrieve key experimental metrics, such as slope temperature. With traditional analysis methods, these quantities would have to be extracted from data using a time-intensive and manual analysis. This framework was developed for a specific diagnostic, but may be applicable to a wide variety of diagnostics common to laser experiments and thus will be especially crucial to the development of high-repetition rate (HRR) diagnostics for HRR laser systems that are coming online.

Highlights

The development of high-intensity high-repetition rate (HRR) lasers, with intensities > 1018 W/cm2 that can operate on the order of 1 Hz or faster, is a quickly maturing field, and a number of laser facilities around the world are already operating in this regime
We describe the development of automated data analysis based on deep neural networks for a synthetic step-filter-based x-ray spectrometer, called the Livermore Tantalum Step Filter (LTSF)
Given to the simplicity of the LTSF diagnostic, it is a good first candidate to study the feasibility and strengths and weaknesses of a machine-learning based automated analysis. While this framework was developed for a specific x-ray diagnostic, these machine learning techniques are applicable to a variety of diagnostics common on short-pulse laser acceleration experiments

Summary

INTRODUCTION

The development of high-intensity high-repetition rate (HRR) lasers, with intensities > 1018 W/cm that can operate on the order of 1 Hz or faster, is a quickly maturing field, and a number of laser facilities around the world are already operating in this regime. The use of these HRR systems will be significant for laserdriven acceleration research, which relies on these high-intensity lasers impinging on matter to generate beam-like high-energy particle and photon sources. The development of high-intensity high-repetition rate (HRR) lasers, with intensities > 1018 W/cm that can operate on the order of 1 Hz or faster, is a quickly maturing field, and a number of laser facilities around the world are already operating in this regime.1 The use of these HRR systems will be significant for laserdriven acceleration research, which relies on these high-intensity lasers impinging on matter to generate beam-like high-energy particle and photon sources. Given to the simplicity of the LTSF diagnostic, it is a good first candidate to study the feasibility and strengths and weaknesses of a machine-learning based automated analysis While this framework was developed for a specific x-ray diagnostic, these machine learning techniques are applicable to a variety of diagnostics common on short-pulse laser acceleration experiments

LTSF DIAGNOSTIC

DEEP LEARNING BASED AUTOMATED DATA ANALYSIS

Description of LTSF training data

FUTURE WORK

CONCLUSIONS