Miles Lopes

Associate Professor, UC Davis

Department of Statistics

Graduate Group in Applied Mathematics (GGAM)

Office: 4105 Mathematical Sciences Building

Email:  melopes à ucdavis.edu

Curriculum Vitae

(Ph.D. UC Berkeley, Statistics. Advisor: Peter J. Bickel)
      Interests:   High-Dimensional Statistics and Machine Learning
      Current and former PhD students:
      Associate Editor:
  • Annals of Statistics

  • Sankhya A
      Support:
  • DOE grant DE-SC0023490,  co-PI,  "Reliable, Scalable, and Data-efficient Randomized Graph Neural Networks for Neural Combinatorial Optimization with Scientific Applications"

  • NSF grant DMS-1915786,  sole PI,  "Bootstrap Methods in Modern Settings: Inference and Computation"

  • NSF grant DMS-1613218,  sole PI,  "Resampling Methods for High-Dimensional and Large-Scale Data"
      Selected publications:

  • Tracy-Widom, Gaussian, and bootstrap: Approximations for leading eigenvalues in high-dimensional PCA
    N. Dörnemann, M. E. Lopes
    Preprint, 2025 [arxiv]

  • Robust max statistics for high-dimensional inference
    M. Liu, M. E. Lopes
    Preprint, 2024 [arxiv]

  • Testing elliptical models in high dimensions
    S. Wang, M. E. Lopes
    Journal of the American Statistical Association, 2025 [link] [arxiv]

  • Randomized numerical linear algebra: A perspective on the field with an eye to software
    with R. Murray, J. Demmel, M. W. Mahoney, et al.
    Preprint, 2023 [arxiv]
    See also related video on the Mutual Information YouTube channel.

  • Improved rates of bootstrap approximation for the operator norm: A coordinate-free approach
    M. E. Lopes
    Annales de l'Institut Henri Poincaré (B) Probabilités et Statistiques, 2025 [pdf] [link]

  • Error estimation for random Fourier features
    J. Yao, N. B. Erichson, M. E. Lopes
    International Conference on Artificial Intelligence and Statistics (AISTATS), 2023 [link] [arxiv]
    (included in oral presentations, top 1.9% of submissions)

  • Central limit theorem and bootstrap approximation in high dimensions: Near 1/√n rates via implicit smoothing
    M. E. Lopes
    Annals of Statistics, 2022 [pdf] [link]

  • Bootstrapping max statistics in high dimensions:
    Near-parametric rates under weak variance decay and application to functional and multinomial data
    M. E. Lopes, Z. Lin, H.-G. Mueller
    Annals of Statistics, 2020 [pdf] [link]

  • Estimating the algorithmic variance of randomized ensembles via the bootstrap
    M. E. Lopes
    Annals of Statistics, 2019 [pdf] [link]

  • Bootstrapping spectral statistics in high dimensions
    M. E. Lopes, A. Blandino, A. Aue
    Biometrika, 2019 [link] [arxiv]

    [extended list]