An adaptive approach to machine learning for compact particle accelerators
Abstract 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 whi...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2021-09-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-021-98785-0 |
| _version_ | 1818742708723777536 |
|---|---|
| author | Alexander Scheinker Frederick Cropp Sergio Paiagua Daniele Filippetto |
| author_facet | Alexander Scheinker Frederick Cropp Sergio Paiagua Daniele Filippetto |
| author_sort | Alexander Scheinker |
| collection | DOAJ |
| description | Abstract 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. |
| first_indexed | 2024-12-18T02:16:49Z |
| format | Article |
| id | doaj.art-1b5bd259315f456690de4a256e722122 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-12-18T02:16:49Z |
| publishDate | 2021-09-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-1b5bd259315f456690de4a256e7221222022-12-21T21:24:22ZengNature PortfolioScientific Reports2045-23222021-09-0111111110.1038/s41598-021-98785-0An adaptive approach to machine learning for compact particle acceleratorsAlexander Scheinker0Frederick Cropp1Sergio Paiagua2Daniele Filippetto3Applied Electrodynamics Group, Los Alamos National LaboratoryDepartment of Physics and Astronomy, University of California Los AngelesLawrence Berkeley National LaboratoryLawrence Berkeley National LaboratoryAbstract 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.https://doi.org/10.1038/s41598-021-98785-0 |
| spellingShingle | Alexander Scheinker Frederick Cropp Sergio Paiagua Daniele Filippetto An adaptive approach to machine learning for compact particle accelerators Scientific Reports |
| title | An adaptive approach to machine learning for compact particle accelerators |
| title_full | An adaptive approach to machine learning for compact particle accelerators |
| title_fullStr | An adaptive approach to machine learning for compact particle accelerators |
| title_full_unstemmed | An adaptive approach to machine learning for compact particle accelerators |
| title_short | An adaptive approach to machine learning for compact particle accelerators |
| title_sort | adaptive approach to machine learning for compact particle accelerators |
| url | https://doi.org/10.1038/s41598-021-98785-0 |
| work_keys_str_mv | AT alexanderscheinker anadaptiveapproachtomachinelearningforcompactparticleaccelerators AT frederickcropp anadaptiveapproachtomachinelearningforcompactparticleaccelerators AT sergiopaiagua anadaptiveapproachtomachinelearningforcompactparticleaccelerators AT danielefilippetto anadaptiveapproachtomachinelearningforcompactparticleaccelerators AT alexanderscheinker adaptiveapproachtomachinelearningforcompactparticleaccelerators AT frederickcropp adaptiveapproachtomachinelearningforcompactparticleaccelerators AT sergiopaiagua adaptiveapproachtomachinelearningforcompactparticleaccelerators AT danielefilippetto adaptiveapproachtomachinelearningforcompactparticleaccelerators |