| Paper | Title | Page |
|---|---|---|
| MOPA55 | Facilitating Machine Learning Collaborations Between Labs, Universities, and Industry | 164 |
|
||
| It is clear from numerous recent community reports, papers, and proposals that machine learning is of tremendous interest for particle accelerator applications. The quickly evolving landscape continues to grow in both the breadth and depth of applications including physics modeling, anomaly detection, controls, diagnostics, and analysis. Consequently, laboratories, universities, and companies across the globe have established dedicated machine learning (ML) and data-science efforts aiming to make use of these new state-of-the-art tools. The current funding environment in the U.S. is structured in a way that supports specific application spaces rather than larger collaboration on community software. Here, we discuss the existing collaboration bottlenecks and how a shift in the funding environment, and how we develop collaborative tools, can help fuel the next wave of ML advancements for particle accelerators. | ||
| DOI • | reference for this paper ※ doi:10.18429/JACoW-NAPAC2022-MOPA55 | |
| About • | Received ※ 10 August 2022 — Revised ※ 11 August 2022 — Accepted ※ 22 August 2022 — Issue date ※ 01 September 2022 | |
| Cite • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |
TUZG1 |
Magnets to ML to Light Sources: Designing from the Browser with Sirepo | |
|
||
| Join us for an afternoon intensive with the Sirepo CAE & Design platform! Experts will present their work and lead tutorials using Sirepo’s apps and tools for magnet design, ML techniques in control systems, and X-ray beamline modeling. Bring your own laptop and pre-register to use the free gateway at Sirepo.com for practical exercises. Dr. Edelen will provide a tutorial on how to design dipoles and undulators using parameterized magnets. Dr. Roussel will demonstrate combining the classical Preisach model of hysteresis with ML techniques to efficiently create non-parametric, high-fidelity models of arbitrary systems exhibiting hysteresis. Also shown will be how using these joint hysteresis-beam models allows users to overcome optimization performance limitations when hysteresis effects are ignored. Dr. Nash will review the capabilities of Shadow and SRW within Sirepo, discuss additional optics tools in the interface including brightness and linear optics, demonstrate translation between the two codes, and show how to continue your simulation work in a python based Jupyter notebook. | ||
| Cite • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |
| FRXE1 | Bayesian Algorithms for Practical Accelerator Control and Adaptive Machine Learning for Time-Varying Systems | 921 |
|
||
| Particle accelerators are complicated machines with thousands of coupled time varying components. The electromagnetic fields of accelerator devices such as magnets and RF cavities drift and are uncertain due to external disturbances, vibrations, temperature changes, and hysteresis. Accelerated charged particle beams are complex objects with 6D phase space dynamics governed by collective effects such as space charge forces, coherent synchrotron radiation, and whose initial phase space distributions change in unexpected and difficult to measure ways. This two-part tutorial presents recent developments in Bayesian methods and adaptive machine learning (ML) techniques for accelerators. Part 1: We introduce Bayesian control algorithms, and we describe how these algorithms can be customized to solve practical accelerator specific problems, including online characterization and optimization. Part 2: We give an overview of adaptive ML (AML) combining adaptive model-independent feedback within physics-informed ML architectures to make ML tools robust to time-variation (distribution shift) and to enable their use further beyond the span of the training data without relying on re-training. | ||
|
Slides FRXE1 [34.283 MB] | |
| DOI • | reference for this paper ※ doi:10.18429/JACoW-NAPAC2022-FRXE1 | |
| About • | Received ※ 08 August 2022 — Revised ※ 10 August 2022 — Accepted ※ 12 August 2022 — Issue date ※ 27 September 2022 | |
| Cite • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |
FRXE2 |
Bayesian Algorithms for Practical Accelerator Control and Adaptive Machine Learning for Time-Varying Systems | |
|
||
| Particle accelerators are complicated machines with thousands of coupled time varying components. The electromagnetic fields of accelerator devices such as magnets and RF cavities drift and are uncertain due to external disturbances, vibrations, temperature changes, and hysteresis. Accelerated charged particle beams are complex objects with 6D phase space dynamics governed by collective effects such as space charge forces, coherent synchrotron radiation, and whose initial phase space distributions change in unexpected and difficult to measure ways. This two-part tutorial presents recent developments in Bayesian methods and adaptive machine learning (ML) techniques for accelerators. Part 1: We introduce Bayesian control algorithms, and we describe how these algorithms can be customized to solve practical accelerator specific problems, including online characterization and optimization. Part 2: We give an overview of adaptive ML (AML) combining adaptive model-independent feedback within physics-informed ML architectures to make ML tools robust to time-variation (distribution shift) and to enable their use further beyond the span of the training data without relying on re-training. | ||
| Cite • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |