Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
134 tokens/sec
GPT-4o
9 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Distance Preserving Machine Learning for Uncertainty Aware Accelerator Capacitance Predictions (2307.02367v1)

Published 5 Jul 2023 in cs.LG and physics.acc-ph

Abstract: Providing accurate uncertainty estimations is essential for producing reliable machine learning models, especially in safety-critical applications such as accelerator systems. Gaussian process models are generally regarded as the gold standard method for this task, but they can struggle with large, high-dimensional datasets. Combining deep neural networks with Gaussian process approximation techniques have shown promising results, but dimensionality reduction through standard deep neural network layers is not guaranteed to maintain the distance information necessary for Gaussian process models. We build on previous work by comparing the use of the singular value decomposition against a spectral-normalized dense layer as a feature extractor for a deep neural Gaussian process approximation model and apply it to a capacitance prediction problem for the High Voltage Converter Modulators in the Oak Ridge Spallation Neutron Source. Our model shows improved distance preservation and predicts in-distribution capacitance values with less than 1% error.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (20)
  1. A review of uncertainty quantification in deep learning: Techniques, applications and challenges. Information Fusion 76, 243–297.
  2. Analog Devices, 2023. LTspice information center. URL: https://www.analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html.
  3. Low-rank incremental methods for computing dominant singular subspaces. Linear Algebra and its Applications 436, 2866–2888.
  4. Uncertainty aware anomaly detection to predict errant beam pulses in the oak ridge spallation neutron source accelerator. Phys. Rev. Accel. Beams 25, 122802. URL: https://link.aps.org/doi/10.1103/PhysRevAccelBeams.25.122802, doi:10.1103/PhysRevAccelBeams.25.122802.
  5. Lifetime testing of metallized thin film capacitors for inverter applications, in: 2013 IEEE 39th Photovoltaic Specialists Conference (PVSC), pp. 3340–3342. doi:10.1109/PVSC.2013.6745166.
  6. Is MC dropout bayesian? arXiv preprint arXiv:2110.04286 .
  7. Dropout as a bayesian approximation: Representing model uncertainty in deep learning, in: international conference on machine learning, PMLR. pp. 1050–1059.
  8. What uncertainties do we need in bayesian deep learning for computer vision? Advances in neural information processing systems 30.
  9. Quantile regression. volume 38. Cambridge university press.
  10. Simple and principled uncertainty estimation with deterministic deep learning via distance awareness. Advances in Neural Information Processing Systems 33, 7498–7512.
  11. Metric learning and manifolds: Preserving the intrinsic geometry. Preprint Department of Statistics, University of Washington .
  12. Progress on Machine Learning for the SNS High Voltage Converter Modulators, in: Proc. 5th Int. Particle Accel. Conf. (NAPAC’22), JACoW Publishing, Geneva, Switzerland. pp. 715–718. URL: https://jacow.org/napac2022/papers/wepa38.pdf, doi:10.18429/JACoW-NAPAC2022-WEPA38.
  13. Random features for large-scale kernel machines. Advances in neural information processing systems 20.
  14. Gaussian Processes for Machine Learning. Adaptive Computation and Machine Learning series, MIT Press. URL: https://gaussianprocess.org/gpml/chapters/RW.pdf.
  15. Design, status, and first operations of the spallation neutron source polyphase resonant converter modulator system, in: Proceedings of the 2003 Particle Accelerator Conference, pp. 553–557 Vol.1. doi:10.1109/PAC.2003.1288975.
  16. Idempotent one-sided approximation of median smoothers. Journal of Approximation Theory 58, 151–163.
  17. Uncertainty aware machine-learning-based surrogate models for particle accelerators: Study at the fermilab booster accelerator complex. Phys. Rev. Accel. Beams 26, 044602. URL: https://link.aps.org/doi/10.1103/PhysRevAccelBeams.26.044602, doi:10.1103/PhysRevAccelBeams.26.044602.
  18. A global geometric framework for nonlinear dimensionality reduction. science 290, 2319–2323.
  19. On feature collapse and deep kernel learning for single forward pass uncertainty. arXiv preprint arXiv:2102.11409 .
  20. Uncertainty estimation using a single deep deterministic neural network, in: International conference on machine learning, PMLR. pp. 9690–9700.
Citations (1)

Summary

We haven't generated a summary for this paper yet.