ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1704.04163
8
36

Spectrum Approximation Beyond Fast Matrix Multiplication: Algorithms and Hardness

13 April 2017
Cameron Musco
Praneeth Netrapalli
Aaron Sidford
Shashanka Ubaru
David P. Woodruff
ArXivPDFHTML
Abstract

Understanding the singular value spectrum of a matrix A∈Rn×nA \in \mathbb{R}^{n \times n}A∈Rn×n is a fundamental task in countless applications. In matrix multiplication time, it is possible to perform a full SVD and directly compute the singular values σ1,...,σn\sigma_1,...,\sigma_nσ1​,...,σn​. However, little is known about algorithms that break this runtime barrier. Using tools from stochastic trace estimation, polynomial approximation, and fast system solvers, we show how to efficiently isolate different ranges of AAA's spectrum and approximate the number of singular values in these ranges. We thus effectively compute a histogram of the spectrum, which can stand in for the true singular values in many applications. We use this primitive to give the first algorithms for approximating a wide class of symmetric matrix norms in faster than matrix multiplication time. For example, we give a (1+ϵ)(1 + \epsilon)(1+ϵ) approximation algorithm for the Schatten-111 norm (the nuclear norm) running in just O~((nnz(A)n1/3+n2)ϵ−3)\tilde O((nnz(A)n^{1/3} + n^2)\epsilon^{-3})O~((nnz(A)n1/3+n2)ϵ−3) time for AAA with uniform row sparsity or O~(n2.18ϵ−3)\tilde O(n^{2.18} \epsilon^{-3})O~(n2.18ϵ−3) time for dense matrices. The runtime scales smoothly for general Schatten-ppp norms, notably becoming O~(p⋅nnz(A)ϵ−3)\tilde O (p \cdot nnz(A) \epsilon^{-3})O~(p⋅nnz(A)ϵ−3) for any p≥2p \ge 2p≥2. At the same time, we show that the complexity of spectrum approximation is inherently tied to fast matrix multiplication in the small ϵ\epsilonϵ regime. We prove that achieving milder ϵ\epsilonϵ dependencies in our algorithms would imply faster than matrix multiplication time triangle detection for general graphs. This further implies that highly accurate algorithms running in subcubic time yield subcubic time matrix multiplication. As an application of our bounds, we show that precisely computing all effective resistances in a graph in less than matrix multiplication time is likely difficult, barring a major algorithmic breakthrough.

View on arXiv
Comments on this paper