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


Current browse context:


Change to browse by:

References & Citations


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

Statistics > Machine Learning

Title: Generalization Bounds for Sparse Random Feature Expansions

Abstract: Random feature methods have been successful in various machine learning tasks, are easy to compute, and come with theoretical accuracy bounds. They serve as an alternative approach to standard neural networks since they can represent similar function spaces without a costly training phase. However, for accuracy, random feature methods require more measurements than trainable parameters, limiting their use for data-scarce applications or problems in scientific machine learning. This paper introduces the sparse random feature expansion to obtain parsimonious random feature models. Specifically, we leverage ideas from compressive sensing to generate random feature expansions with theoretical guarantees even in the data-scarce setting. In particular, we provide generalization bounds for functions in a certain class (that is dense in a reproducing kernel Hilbert space) depending on the number of samples and the distribution of features. The generalization bounds improve with additional structural conditions, such as coordinate sparsity, compact clusters of the spectrum, or rapid spectral decay. In particular, by introducing sparse features, i.e. features with random sparse weights, we provide improved bounds for low order functions. We show that the sparse random feature expansions outperforms shallow networks in several scientific machine learning tasks.
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Numerical Analysis (math.NA); Optimization and Control (math.OC); Probability (math.PR)
Cite as: arXiv:2103.03191 [stat.ML]
  (or arXiv:2103.03191v3 [stat.ML] for this version)

Submission history

From: Hayden Schaeffer [view email]
[v1] Thu, 4 Mar 2021 17:53:54 GMT (677kb,D)
[v2] Mon, 19 Apr 2021 20:44:43 GMT (1370kb,D)
[v3] Fri, 20 Aug 2021 18:00:39 GMT (1209kb,D)

Link back to: arXiv, form interface, contact.