We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: A Random Matrix Analysis of Random Fourier Features: Beyond the Gaussian Kernel, a Precise Phase Transition, and the Corresponding Double Descent

Abstract: This article characterizes the exact asymptotics of random Fourier feature (RFF) regression, in the realistic setting where the number of data samples $n$, their dimension $p$, and the dimension of feature space $N$ are all large and comparable. In this regime, the random RFF Gram matrix no longer converges to the well-known limiting Gaussian kernel matrix (as it does when $N \to \infty$ alone), but it still has a tractable behavior that is captured by our analysis. This analysis also provides accurate estimates of training and test regression errors for large $n,p,N$. Based on these estimates, a precise characterization of two qualitatively different phases of learning, including the phase transition between them, is provided; and the corresponding double descent test error curve is derived from this phase transition behavior. These results do not depend on strong assumptions on the data distribution, and they perfectly match empirical results on real-world data sets.
Comments: NeurIPS 2020
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Journal reference: Advances in Neural Information Processing Systems (NeurIPS 2020) 33:13939-13950; Journal of Statistical Mechanics: Theory and Experiment, Volume 2021, December 2021
DOI: 10.1088/1742-5468/ac3a77
Cite as: arXiv:2006.05013 [stat.ML]
  (or arXiv:2006.05013v2 [stat.ML] for this version)

Submission history

From: Zhenyu Liao [view email]
[v1] Tue, 9 Jun 2020 02:05:40 GMT (52kb)
[v2] Sat, 19 Dec 2020 06:28:07 GMT (58kb)

Link back to: arXiv, form interface, contact.