A machine learning model was created to predict the electron spectrum generated by a GeV-class laser wakefield accelerator.The model was constructed from variational convolutional neural networks,which mapped the resu...A machine learning model was created to predict the electron spectrum generated by a GeV-class laser wakefield accelerator.The model was constructed from variational convolutional neural networks,which mapped the results of secondary laser and plasma diagnostics to the generated electron spectrum.An ensemble of trained networks was used to predict the electron spectrum and to provide an estimation of the uncertainty of that prediction.It is anticipated that this approach will be useful for inferring the electron spectrum prior to undergoing any process that can alter or destroy the beam.In addition,the model provides insight into the scaling of electron beam properties due to stochastic fluctuations in the laser energy and plasma electron density.展开更多
The interaction of relativistically intense lasers with opaque targets represents a highly non-linear,multi-dimensional parameter space.This limits the utility of sequential 1D scanning of experimental parameters for ...The interaction of relativistically intense lasers with opaque targets represents a highly non-linear,multi-dimensional parameter space.This limits the utility of sequential 1D scanning of experimental parameters for the optimization of secondary radiation,although to-date this has been the accepted methodology due to low data acquisition rates.High repetition-rate(HRR)lasers augmented by machine learning present a valuable opportunity for efficient source optimization.Here,an automated,HRR-compatible system produced high-fidelity parameter scans,revealing the influence of laser intensity on target pre-heating and proton generation.A closed-loop Bayesian optimization of maximum proton energy,through control of the laser wavefront and target position,produced proton beams with equivalent maximum energy to manually optimized laser pulses but using only 60%of the laser energy.This demonstration of automated optimization of laser-driven proton beams is a crucial step towards deeper physical insight and the construction of future radiation sources.展开更多
基金supported by UK STFC ST/V001639/1,UK EPSRC EP/V049577/1 and EP/V044397/1Horizon 2020 funding under European Research Council(ERC)Grant Agreement No.682399+1 种基金support from the Royal Society URF-R1221874support from US DOE grant DESC0016804
文摘A machine learning model was created to predict the electron spectrum generated by a GeV-class laser wakefield accelerator.The model was constructed from variational convolutional neural networks,which mapped the results of secondary laser and plasma diagnostics to the generated electron spectrum.An ensemble of trained networks was used to predict the electron spectrum and to provide an estimation of the uncertainty of that prediction.It is anticipated that this approach will be useful for inferring the electron spectrum prior to undergoing any process that can alter or destroy the beam.In addition,the model provides insight into the scaling of electron beam properties due to stochastic fluctuations in the laser energy and plasma electron density.
基金support from the UK STFC grants ST/V001639/1 with the XFEL Physical Sciences Hub and ST/P002021/1the UK EPSRC grants EP/V049577/1 and EP/R006202/1+5 种基金as well as the U.S.DOE Office of Science,Fusion Energy Sciences under FWP No.100182in part by the National Science Foundation under Grant No.1632708 and Award No.PHY–1903414M.J.V.S.acknowledges support from the Royal Society URFR1221874support from the DOE NNSA SSGF program under DE-NA0003960support from the U.S.DOE grant DESC0016804support from the project‘Advanced research using high-intensity laser-produced photons and particles’(CZ.02.1.01/0.0/0.0/16_019/0000789)from the European Regional Development Fund(ADONIS)。
文摘The interaction of relativistically intense lasers with opaque targets represents a highly non-linear,multi-dimensional parameter space.This limits the utility of sequential 1D scanning of experimental parameters for the optimization of secondary radiation,although to-date this has been the accepted methodology due to low data acquisition rates.High repetition-rate(HRR)lasers augmented by machine learning present a valuable opportunity for efficient source optimization.Here,an automated,HRR-compatible system produced high-fidelity parameter scans,revealing the influence of laser intensity on target pre-heating and proton generation.A closed-loop Bayesian optimization of maximum proton energy,through control of the laser wavefront and target position,produced proton beams with equivalent maximum energy to manually optimized laser pulses but using only 60%of the laser energy.This demonstration of automated optimization of laser-driven proton beams is a crucial step towards deeper physical insight and the construction of future radiation sources.
