Temporal power reconstruction for an x-ray free-electron laser using convolutional neural networks

Ren, Xiaoming; Edelen, Auralee; Lutman, Alberto; Marcus, Gabriel; Maxwell, Timothy; Ratner, Daniel

doi:10.1103/physrevaccelbeams.23.040701

“…Neural networks are trained using the minibatch stochastic gradient decent optimization algorithm [13] driven by a loss function. For most regression problems, the choice of the loss function defaults to the mean squared error (MSE) [4][5][6]17]. However, a MSE loss function treats pixels as uncorrelated features and is found to result in overly smoothed images and loss of high-frequency features in high-resolution image-generation applications [21].…”

Section: B Loss Functionmentioning

confidence: 99%

“…The operation of large-scale scientific user facilities such as the European XFEL [1] is very challenging, as it is necessary to meet the specifications of various user experiments [2] and to be capable of switching the machine status rapidly. Machine learning, especially deep learning, has provided powerful tools for accelerator physicists to build fast-prediction surrogate models [3][4][5] and to extract essential information [6][7][8] from large amounts of data in recent years. These machine-learning models can be extremely useful for building virtual accelerators, which are capable of making fast predictions of the behavior of beams [9], assisting accelerator tuning by virtually bringing destructive diagnostics online [4], providing an initial guess of input parameters for model-independent adaptive feedback control algorithms [10,11], and driving modelbased feedback control algorithms [12].…”

Section: Introductionmentioning

confidence: 99%

High-Fidelity Prediction of Megapixel Longitudinal Phase-Space Images of Electron Beams Using Encoder-Decoder Neural Networks

Zhu

¹

,

Chen²,

Brinker³

et al. 2021

View full text Add to dashboard Cite

Modeling of large-scale research facilities is extremely challenging due to complex physical processes and engineering problems. Here, we adopt a data-driven approach to model the longitudinal phase-spacediagnostic beamline at the photoinector of the European XFEL with an encoder-decoder neural-network model. A deep convolutional neural network (decoder) is used to build images, measured on the screen, from a small feature map generated by another neural network (encoder). We demonstrate that the model, trained only with experimental data, can make high-fidelity predictions of megapixel images for the longitudinal phase-space measurement without any prior knowledge of photoinjectors or electron beams. The prediction significantly outperforms existing methods. We also show the scalability and interpretability of the model by sharing the same decoder with more than one encoder, used for different setups of the photoinjector, and propose a pragmatic way to model a facility with various diagnostics and working points. This opens the door to a way of accurately modeling a photoinjector using neural networks and experimental data. The approach can possibly be extended to the whole accelerator and even other types of scientific facility.

show abstract

“…To give a few concrete examples, various ML methods have now been demonstrated for a wide range of systems such as molecular and materials science studies 3 , for use in optical communications and photonics 4 , to accurately predict battery life 5 , to accelerate lattice Monte Carlo simulations using neural networks 6 , for studying complex networks 7 , for characterizing surface microstructure of complex materials 8 , for chemical discovery 9 , for noninvasive identification of Hypotension using convolutional-deconvolutional networks 10 , for active matter analysis by using deep neural networks to track objects 11 , for imputation of missing physiological waveform data by using convolutional autoencoders 12 , for optimizing operational problems in hospitals 13 , for cardiovascular disease risk prediction 14 , for particle physics 15 , for antimicrobial studies 16 , for pattern recognition for optical microscopy images of metallurgical microstructures 17 , for learning Perovskit bandgaps 18 , for real-time mapping of electron backscatter diffraction (EBSD) patterns to crystal orientations 19 , for speeding up simulation-based accelerator optimization studies 20 , for Bayesian optimization of free electron lasers (FEL) 21 , for temporal power reconstruction of FELs 22 , for various applications at the Large Hadron Collider (LHC) at CERN including optics corrections and detecting faulty beam position monitors [23][24][25][26] , for reconstruction of a storage ring's linear optics based on Bayesian inference 27 , to analyze beam position monitor placement in accelerators to find arrangements with the lowest probable predictive errors based on Bayesian Gaussian regression 28 , for temporal shaping of electron bunches in particle accelerators 29 , for stabilization of source properties in synchrotron light sources 30 , and to represent many-body interactions with restricted-Boltzmann-machine neural networks 31 .…”

mentioning

confidence: 99%

An investigation of aqueous ammonium nitrate aerosols with soft X-ray spectroscopy

Weeraratna

¹

,

Kostko

²

,

Ahmed

³

2021

View full text Add to dashboard Cite

Machine learning (ML) tools are able to learn relationships between the inputs and outputs of large complex systems directly from data. However, for time-varying systems, the predictive capabilities of ML tools degrade if the systems are no longer accurately represented by the data with which the ML models were trained. For complex systems, re-training is only possible if the changes are slow relative to the rate at which large numbers of new input-output training data can be non-invasively recorded. In this work, we present an approach to deep learning for time-varying systems that does not require re-training, but uses instead an adaptive feedback in the architecture of deep convolutional neural networks (CNN). The feedback is based only on available system output measurements and is applied in the encoded low-dimensional dense layers of the encoder-decoder CNNs. First, we develop an inverse model of a complex accelerator system to map output beam measurements to input beam distributions, while both the accelerator components and the unknown input beam distribution vary rapidly with time. We then demonstrate our method on experimental measurements of the input and output beam distributions of the HiRES ultra-fast electron diffraction (UED) beam line at Lawrence Berkeley National Laboratory, and showcase its ability for automatic tracking of the time varying photocathode quantum efficiency map. Our method can be successfully used to aid both physics and ML-based surrogate online models to provide non-invasive beam diagnostics.

show abstract

“…To give a few concrete examples, various ML methods have now been demonstrated for a wide range of systems such as molecular and materials science studies 3 , for use in optical communications and photonics 4 , to accurately predict battery life 5 , to accelerate lattice Monte Carlo simulations using neural networks 6 , for studying complex networks 7 , for characterizing surface microstructure of complex materials 8 , for chemical discovery 9 , for noninvasive identification of Hypotension using convolutional-deconvolutional networks 10 , for active matter analysis by using deep neural networks to track objects 11 , for imputation of missing physiological waveform data by using convolutional autoencoders 12 , for optimizing operational problems in hospitals 13 , for cardiovascular disease risk prediction 14 , for particle physics 15 , for antimicrobial studies 16 , for pattern recognition for optical microscopy images of metallurgical microstructures 17 , for learning Perovskit bandgaps 18 , for real-time mapping of electron backscatter diffraction (EBSD) patterns to crystal orientations 19 , for speeding up simulation-based accelerator optimization studies 20 , for Bayesian optimization of free electron lasers (FEL) 21 , for temporal power reconstruction of FELs 22 , for various applications at the Large Hadron Collider (LHC) at CERN including optics corrections and detecting faulty beam position monitors 23 – 26 , for reconstruction of a storage ring’s linear optics based on Bayesian inference 27 , to analyze beam position monitor placement in accelerators to find arrangements with the lowest probable predictive errors based on Bayesian Gaussian regression 28 , for temporal shaping of electron bunches in particle accelerators 29 , for stabilization of source properties in synchrotron light sources 30 , and to represent many-body interactions with restricted-Boltzmann-machine neural networks 31 .…”

Section: Introductionmentioning

confidence: 99%

An adaptive approach to machine learning for compact particle accelerators

Scheinker

¹

,

Cropp

²

,

Paiagua

³

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Machine learning (ML) tools are able to learn relationships between the inputs and outputs of large complex systems directly from data. However, for time-varying systems, the predictive capabilities of ML tools degrade if the systems are no longer accurately represented by the data with which the ML models were trained. For complex systems, re-training is only possible if the changes are slow relative to the rate at which large numbers of new input-output training data can be non-invasively recorded. In this work, we present an approach to deep learning for time-varying systems that does not require re-training, but uses instead an adaptive feedback in the architecture of deep convolutional neural networks (CNN). The feedback is based only on available system output measurements and is applied in the encoded low-dimensional dense layers of the encoder-decoder CNNs. First, we develop an inverse model of a complex accelerator system to map output beam measurements to input beam distributions, while both the accelerator components and the unknown input beam distribution vary rapidly with time. We then demonstrate our method on experimental measurements of the input and output beam distributions of the HiRES ultra-fast electron diffraction (UED) beam line at Lawrence Berkeley National Laboratory, and showcase its ability for automatic tracking of the time varying photocathode quantum efficiency map. Our method can be successfully used to aid both physics and ML-based surrogate online models to provide non-invasive beam diagnostics.

show abstract

Temporal power reconstruction for an x-ray free-electron laser using convolutional neural networks

Cited by 14 publications

References 33 publications

High-Fidelity Prediction of Megapixel Longitudinal Phase-Space Images of Electron Beams Using Encoder-Decoder Neural Networks

High-Fidelity Prediction of Megapixel Longitudinal Phase-Space Images of Electron Beams Using Encoder-Decoder Neural Networks

An investigation of aqueous ammonium nitrate aerosols with soft X-ray spectroscopy

An adaptive approach to machine learning for compact particle accelerators

Contact Info

Product

Resources

About