Beyond optimization -- supervised learning applications in relativistic laser-plasma experiments
TL;DR: In this article, the beam charge of electrons produced in a laser wakefield accelerator given the laser wavefront change caused by a deformable mirror is predicted using machine learning techniques beyond optimization purposes.
Abstract: We explore the applications of machine learning techniques in relativistic laser-plasma experiments beyond optimization purposes. We predict the beam charge of electrons produced in a laser wakefield accelerator given the laser wavefront change caused by a deformable mirror. Machine learning enables feature analysis beyond merely searching for an optimal beam charge, showing that specific aberrations in the laser wavefront are favored in generating higher beam charges. Supervised learning models allow characterizing the measured data quality as well as recognizing irreproducible data and potential outliers. We also include virtual measurement errors in the experimental data to examine the model robustness under these conditions. This work demonstrates how machine learning methods can benefit data analysis and physics interpretation in a highly nonlinear problem of relativistic laser-plasma interaction.
Citations
More filters
TL;DR: In this article , the authors present an overview of relevant machine learning methods with focus on applicability to laser-plasma physics, including its important sub-fields of laserplasma acceleration and inertial confinement fusion.
Abstract: Laser-plasma physics has developed rapidly over the past few decades as high-power lasers have become both increasingly powerful and more widely available. Early experimental and numerical research in this field was restricted to single-shot experiments with limited parameter exploration. However, recent technological improvements make it possible to gather an increasing amount of data, both in experiments and simulations. This has sparked interest in using advanced techniques from mathematics, statistics and computer science to deal with, and benefit from, big data. At the same time, sophisticated modeling techniques also provide new ways for researchers to effectively deal with situations in which still only sparse amounts of data are available. This paper aims to present an overview of relevant machine learning methods with focus on applicability to laser-plasma physics, including its important sub-fields of laser-plasma acceleration and inertial confinement fusion.
4 citations
TL;DR: In this paper , a state-of-the-art object detection architecture was fine-tuned to work on a variety of object detection tasks in a high-power laser laboratory.
Abstract: Abstract The recent advent of deep artificial neural networks has resulted in a dramatic increase in performance for object classification and detection. While pre-trained with everyday objects, we find that a state-of-the-art object detection architecture can very efficiently be fine-tuned to work on a variety of object detection tasks in a high-power laser laboratory. In this paper, three exemplary applications are presented. We show that the plasma waves in a laser–plasma accelerator can be detected and located on the optical shadowgrams. The plasma wavelength and plasma density are estimated accordingly. Furthermore, we present the detection of all the peaks in an electron energy spectrum of the accelerated electron beam, and the beam charge of each peak is estimated accordingly. Lastly, we demonstrate the detection of optical damage in a high-power laser system. The reliability of the object detector is demonstrated over 1000 laser shots in each application. Our study shows that deep object detection networks are suitable to assist online and offline experimental analysis, even with small training sets. We believe that the presented methodology is adaptable yet robust, and we encourage further applications in Hz-level or kHz-level high-power laser facilities regarding the control and diagnostic tools, especially for those involving image data.
2 citations
TL;DR: In this paper, the authors report on the development of ML-based diagnostics for experiments on high-intensity laser-matter interactions and test the application of principal component analysis, data augmentation and training with data that has superimposed noise of gradually increasing amplitude.
Abstract: The power of machine learning (ML) in feature identification can be harnessed for determining quantities in experiments that are difficult to measure directly. However, if an ML model is trained on simulated data, rather than experimental results, the differences between the two can pose an obstacle to reliable data extraction. Here we report on the development of ML-based diagnostics for experiments on high-intensity laser–matter interactions. With the intention to accentuate robust, physics-governed features, the presence of which is tolerant to such differences, we test the application of principal component analysis, data augmentation and training with data that has superimposed noise of gradually increasing amplitude. Using synthetic data of simulated experiments, we identify that the approach based on the noise of increasing amplitude yields the most accurate ML models and thus is likely to be useful in similar projects on ML-based diagnostics.
1 citations
References
More filters
TL;DR: It is shown that standard multilayer feedforward networks with as few as a single hidden layer and arbitrary bounded and nonconstant activation function are universal approximators with respect to L p (μ) performance criteria, for arbitrary finite input environment measures μ.
5,593 citations
TL;DR: In this paper, an intense electromagnetic pulse can create a weak of plasma oscillations through the action of the nonlinear ponderomotive force, and electrons trapped in the wake can be accelerated to high energy.
Abstract: An intense electromagnetic pulse can create a weak of plasma oscillations through the action of the nonlinear ponderomotive force. Electrons trapped in the wake can be accelerated to high energy. Existing glass lasers of power density ${10}^{18}$W/${\mathrm{cm}}^{2}$ shone on plasmas of densities ${10}^{18}$ ${\mathrm{cm}}^{\ensuremath{-}3}$ can yield gigaelectronvolts of electron energy per centimeter of acceleration distance. This acceleration mechanism is demonstrated through computer simulation. Applications to accelerators and pulsers are examined.
3,867 citations
TL;DR: In this paper, the basic physics of laser pulse evolution in underdense plasmas is also reviewed, including the propagation, self-focusing, and guiding of laser pulses in uniform density channels and with preformed density channels.
Abstract: Laser-driven plasma-based accelerators, which are capable of supporting fields in excess of 100 GV/m, are reviewed. This includes the laser wakefield accelerator, the plasma beat wave accelerator, the self-modulated laser wakefield accelerator, plasma waves driven by multiple laser pulses, and highly nonlinear regimes. The properties of linear and nonlinear plasma waves are discussed, as well as electron acceleration in plasma waves. Methods for injecting and trapping plasma electrons in plasma waves are also discussed. Limits to the electron energy gain are summarized, including laser pulse diffraction, electron dephasing, laser pulse energy depletion, and beam loading limitations. The basic physics of laser pulse evolution in underdense plasmas is also reviewed. This includes the propagation, self-focusing, and guiding of laser pulses in uniform plasmas and with preformed density channels. Instabilities relevant to intense short-pulse laser-plasma interactions, such as Raman, self-modulation, and hose instabilities, are discussed. Experiments demonstrating key physics, such as the production of high-quality electron bunches at energies of 0.1-1 GeV, are summarized.
2,108 citations
TL;DR: It is demonstrated that this randomization of electrons in phase space can be suppressed and that the quality of the electron beams can be dramatically enhanced.
Abstract: Particle accelerators are used in a wide variety of fields, ranging from medicine and biology to high-energy physics. The accelerating fields in conventional accelerators are limited to a few tens of MeV m(-1), owing to material breakdown at the walls of the structure. Thus, the production of energetic particle beams currently requires large-scale accelerators and expensive infrastructures. Laser-plasma accelerators have been proposed as a next generation of compact accelerators because of the huge electric fields they can sustain (>100 GeV m(-1)). However, it has been difficult to use them efficiently for applications because they have produced poor-quality particle beams with large energy spreads, owing to a randomization of electrons in phase space. Here we demonstrate that this randomization can be suppressed and that the quality of the electron beams can be dramatically enhanced. Within a length of 3 mm, the laser drives a plasma bubble that traps and accelerates plasma electrons. The resulting electron beam is extremely collimated and quasi-monoenergetic, with a high charge of 0.5 nC at 170 MeV.
1,854 citations
TL;DR: A laser accelerator that produces electron beams with an energy spread of a few per cent, low emittance and increased energy (more than 109 electrons above 80 MeV) and opens the way for compact and tunable high-brightness sources of electrons and radiation.
Abstract: Laser-driven accelerators, in which particles are accelerated by the electric field of a plasma wave (the wakefield) driven by an intense laser, have demonstrated accelerating electric fields of hundreds of GV m-1 (refs 1–3) These fields are thousands of times greater than those achievable in conventional radio-frequency accelerators, spurring interest in laser accelerators4,5 as compact next-generation sources of energetic electrons and radiation To date, however, acceleration distances have been severely limited by the lack of a controllable method for extending the propagation distance of the focused laser pulse The ensuing short acceleration distance results in low-energy beams with 100 per cent electron energy spread1,2,3, which limits potential applications Here we demonstrate a laser accelerator that produces electron beams with an energy spread of a few per cent, low emittance and increased energy (more than 109 electrons above 80 MeV) Our technique involves the use of a preformed plasma density channel to guide a relativistically intense laser, resulting in a longer propagation distance The results open the way for compact and tunable high-brightness sources of electrons and radiation
1,749 citations