-
Quantifying Uncertainty for Machine Learning Based Diagnostic
Authors:
Owen Convery,
Lewis Smith,
Yarin Gal,
Adi Hanuka
Abstract:
Virtual Diagnostic (VD) is a deep learning tool that can be used to predict a diagnostic output. VDs are especially useful in systems where measuring the output is invasive, limited, costly or runs the risk of damaging the output. Given a prediction, it is necessary to relay how reliable that prediction is. This is known as 'uncertainty quantification' of a prediction. In this paper, we use ensemb…
▽ More
Virtual Diagnostic (VD) is a deep learning tool that can be used to predict a diagnostic output. VDs are especially useful in systems where measuring the output is invasive, limited, costly or runs the risk of damaging the output. Given a prediction, it is necessary to relay how reliable that prediction is. This is known as 'uncertainty quantification' of a prediction. In this paper, we use ensemble methods and quantile regression neural networks to explore different ways of creating and analyzing prediction's uncertainty on experimental data from the Linac Coherent Light Source at SLAC. We aim to accurately and confidently predict the current profile or longitudinal phase space images of the electron beam. The ability to make informed decisions under uncertainty is crucial for reliable deployment of deep learning tools on safety-critical systems as particle accelerators.
△ Less
Submitted 29 July, 2021;
originally announced July 2021.
-
Turn-Key Constrained Parameter Space Exploration for Particle Accelerators Using Bayesian Active Learning
Authors:
Ryan Roussel,
Juan Pablo Gonzalez-Aguilera,
Young-Kee Kim,
Eric Wisniewski,
Wanming Liu,
Philippe Piot,
John Power,
Adi Hanuka,
Auralee Edelen
Abstract:
Particle accelerators are invaluable discovery engines in the chemical, biological and physical sciences. Characterization of the accelerated beam response to accelerator input parameters is of-ten the first step when conducting accelerator-based experiments. Currently used techniques for characterization, such as grid-like parameter sampling scans, become impractical when extended to higher dimen…
▽ More
Particle accelerators are invaluable discovery engines in the chemical, biological and physical sciences. Characterization of the accelerated beam response to accelerator input parameters is of-ten the first step when conducting accelerator-based experiments. Currently used techniques for characterization, such as grid-like parameter sampling scans, become impractical when extended to higher dimensional input spaces, when complicated measurement constraints are present, or prior information is known about the beam response is scarce. In this work, we describe an adaptation of the popular Bayesian optimization algorithm, which enables a turn-key exploration algorithm that replaces parameter scans and minimizes prior information needed about the measurements' behavior and associated measurement constraints. We experimentally demonstrate that our algorithm autonomously conducts an adaptive, multi-parameter exploration of input parameter space,while navigating a highly constrained, single-shot beam phase-space measurement. In addition to applications in accelerator-based scientific experiments, this algorithm addresses challenges shared by many scientific disciplines and is thus applicable to autonomously conducting experiments over a broad range of research topics.
△ Less
Submitted 16 June, 2021;
originally announced June 2021.
-
Uncertainty Quantification for Virtual Diagnostic of Particle Accelerators
Authors:
Owen Convery,
Lewis Smith,
Yarin Gal,
Adi Hanuka
Abstract:
Virtual Diagnostic (VD) is a computational tool based on deep learning that can be used to predict a diagnostic output. VDs are especially useful in systems where measuring the output is invasive, limited, costly or runs the risk of altering the output. Given a prediction, it is necessary to relay how reliable that prediction is, i.e. quantify the uncertainty of the prediction. In this paper, we u…
▽ More
Virtual Diagnostic (VD) is a computational tool based on deep learning that can be used to predict a diagnostic output. VDs are especially useful in systems where measuring the output is invasive, limited, costly or runs the risk of altering the output. Given a prediction, it is necessary to relay how reliable that prediction is, i.e. quantify the uncertainty of the prediction. In this paper, we use ensemble methods and quantile regression neural networks to explore different ways of creating and analyzing prediction's uncertainty on experimental data from the Linac Coherent Light Source at SLAC National Lab. We aim to accurately and confidently predict the current profile or longitudinal phase space images of the electron beam. The ability to make informed decisions under uncertainty is crucial for reliable deployment of deep learning tools on safety-critical systems as particle accelerators.
△ Less
Submitted 10 May, 2021;
originally announced May 2021.
-
Multi-Objective Bayesian Optimization for Accelerator Tuning
Authors:
Ryan Roussel,
Adi Hanuka,
Auralee Edelen
Abstract:
Particle accelerators require constant tuning during operation to meet beam quality, total charge and particle energy requirements for use in a wide variety of physics, chemistry and biology experiments. Maximizing the performance of an accelerator facility often necessitates multi-objective optimization, where operators must balance trade-offs between multiple objectives simultaneously, often usi…
▽ More
Particle accelerators require constant tuning during operation to meet beam quality, total charge and particle energy requirements for use in a wide variety of physics, chemistry and biology experiments. Maximizing the performance of an accelerator facility often necessitates multi-objective optimization, where operators must balance trade-offs between multiple objectives simultaneously, often using limited, temporally expensive beam observations. Usually, accelerator optimization problems are solved offline, prior to actual operation, with advanced beamline simulations and parallelized optimization methods (NSGA-II, Swarm Optimization). Unfortunately, it is not feasible to use these methods for online multi-objective optimization, since beam measurements can only be done in a serial fashion, and these optimization methods require a large number of measurements to converge to a useful solution.Here, we introduce a multi-objective Bayesian optimization scheme, which finds the full Pareto front of an accelerator optimization problem efficiently in a serialized manner and is thus a critical step towards practical online multi-objective optimization in accelerators.This method uses a set of Gaussian process surrogate models, along with a multi-objective acquisition function, which reduces the number of observations needed to converge by at least an order of magnitude over current methods.We demonstrate how this method can be modified to specifically solve optimization challenges posed by the tuning of accelerators.This includes the addition of optimization constraints, objective preferences and costs related to changing accelerator parameters.
△ Less
Submitted 21 April, 2021; v1 submitted 19 October, 2020;
originally announced October 2020.
-
Accurate and confident prediction of electron beam longitudinal properties using spectral virtual diagnostics
Authors:
A. Hanuka,
C. Emma,
T. Maxwell,
A. Fisher,
B. Jacobson,
M. J. Hogan,
Z. Huang
Abstract:
Longitudinal phase space (LPS) provides a critical information about electron beam dynamics for various scientific applications. For example, it can give insight into the high-brightness X-ray radiation from a free electron laser. Existing diagnostics are invasive, and often times cannot operate at the required resolution. In this work we present a machine learning-based Virtual Diagnostic (VD) to…
▽ More
Longitudinal phase space (LPS) provides a critical information about electron beam dynamics for various scientific applications. For example, it can give insight into the high-brightness X-ray radiation from a free electron laser. Existing diagnostics are invasive, and often times cannot operate at the required resolution. In this work we present a machine learning-based Virtual Diagnostic (VD) tool to accurately predict the LPS for every shot using spectral information collected non-destructively from the radiation of relativistic electron beam. We demonstrate the tool's accuracy for three different case studies with experimental or simulated data. For each case, we introduce a method to increase the confidence in the VD tool. We anticipate that spectral VD would improve the setup and understanding of experimental configurations at DOE's user facilities as well as data sorting and analysis. The spectral VD can provide confident knowledge of the longitudinal bunch properties at the next generation of high-repetition rate linear accelerators while reducing the load on data storage, readout and streaming requirements.
△ Less
Submitted 27 September, 2020;
originally announced September 2020.
-
Physics-informed Gaussian Process for Online Optimization of Particle Accelerators
Authors:
Adi Hanuka,
X. Huang,
J. Shtalenkova,
D. Kennedy,
A. Edelen,
V. R. Lalchand,
D. Ratner,
J. Duris
Abstract:
High-dimensional optimization is a critical challenge for operating large-scale scientific facilities. We apply a physics-informed Gaussian process (GP) optimizer to tune a complex system by conducting efficient global search. Typical GP models learn from past observations to make predictions, but this reduces their applicability to new systems where archive data is not available. Instead, here we…
▽ More
High-dimensional optimization is a critical challenge for operating large-scale scientific facilities. We apply a physics-informed Gaussian process (GP) optimizer to tune a complex system by conducting efficient global search. Typical GP models learn from past observations to make predictions, but this reduces their applicability to new systems where archive data is not available. Instead, here we use a fast approximate model from physics simulations to design the GP model. The GP is then employed to make inferences from sequential online observations in order to optimize the system. Simulation and experimental studies were carried out to demonstrate the method for online control of a storage ring. We show that the physics-informed GP outperforms current routinely used online optimizers in terms of convergence speed, and robustness on this task. The ability to inform the machine-learning model with physics may have wide applications in science.
△ Less
Submitted 8 September, 2020;
originally announced September 2020.
-
Online tuning and light source control using a physics-informed Gaussian process Adi
Authors:
A. Hanuka,
J. Duris,
J. Shtalenkova,
D. Kennedy,
A. Edelen,
D. Ratner,
X. Huang
Abstract:
Operating large-scale scientific facilities often requires fast tuning and robust control in a high dimensional space. In this paper we introduce a new physics-informed optimization algorithm based on Gaussian process regression. Our method takes advantage of the existing domain knowledge in the form of realizations of a physics model of the observed system. We have applied a physics-informed Gaus…
▽ More
Operating large-scale scientific facilities often requires fast tuning and robust control in a high dimensional space. In this paper we introduce a new physics-informed optimization algorithm based on Gaussian process regression. Our method takes advantage of the existing domain knowledge in the form of realizations of a physics model of the observed system. We have applied a physics-informed Gaussian Process method experimentally at the SPEAR3 storage ring to demonstrate online accelerator optimization. This method outperforms Gaussian Process trained on data as well as the standard approach routinely used for operation, in terms of convergence speed and optimal point. The proposed method could be applicable to automatic tuning and control of other complex systems, without a prerequisite for any observed data.
△ Less
Submitted 4 November, 2019;
originally announced November 2019.
-
Bayesian optimization of a free-electron laser
Authors:
Joseph Duris,
Dylan Kennedy,
Adi Hanuka,
Jane Shtalenkova,
Auralee Edelen,
Adam Egger,
Tyler Cope,
Daniel Ratner
Abstract:
The Linac Coherent Light Source changes configurations multiple times per day, necessitating fast tuning strategies to reduce setup time for successive experiments. To this end, we employ a Bayesian approach to transport optics tuning to optimize groups of quadrupole magnets. We use a Gaussian process to provide a probabilistic model of the machine response with respect to control parameters from…
▽ More
The Linac Coherent Light Source changes configurations multiple times per day, necessitating fast tuning strategies to reduce setup time for successive experiments. To this end, we employ a Bayesian approach to transport optics tuning to optimize groups of quadrupole magnets. We use a Gaussian process to provide a probabilistic model of the machine response with respect to control parameters from a modest number of samples. Subsequent samples are selected during optimization using a statistical test combining the model prediction and uncertainty. The model parameters are fit from archived scans, and correlations between devices are added from a simple beam transport model. The result is a sample-efficient optimization routine, which we show significantly outperforms existing optimizers.
△ Less
Submitted 12 September, 2019;
originally announced September 2019.
-
Nonlinear response in high-field dielectric laser accelerators
Authors:
D. Cesar,
S. Custodio,
J. Maxson,
P. Musumeci,
X. Shen,
E. Threlkeld,
R. J. England,
A. Hanuka,
I. V. Makasyuk,
E. A. Peralta,
K. P. Wootton,
Z. Wu
Abstract:
Laser powered dielectric structures achieve high-gradient particle acceleration by taking advantage of modern laser technology capable of producing electric fields in excess of 10GV/m. These fields can drive the bulk dielectric beyond its linear response, and break the phase synchronicity between the accelerating field and the electrons. We show how control of the pulse dispersion can be used to c…
▽ More
Laser powered dielectric structures achieve high-gradient particle acceleration by taking advantage of modern laser technology capable of producing electric fields in excess of 10GV/m. These fields can drive the bulk dielectric beyond its linear response, and break the phase synchronicity between the accelerating field and the electrons. We show how control of the pulse dispersion can be used to compensate the effect of self-phase modulation and maximize the energy gain in the laser accelerator.In our experiment, a high brightness 8MeV e-beam is used to probe accelerating fields of 1.8GV/m in a 'grating-reset' dielectric structure illuminated by a 45fs laser pulse with a fluence of 0.7J/cm$^2$.
△ Less
Submitted 7 July, 2017;
originally announced July 2017.
