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. 1603.07377
19
31

Overcoming The Limitations of Phase Transition by Higher Order Analysis of Regularization Techniques

23 March 2016
Haolei Weng
A. Maleki
Le Zheng
ArXivPDFHTML
Abstract

We study the problem of estimating β∈Rp\beta \in \mathbb{R}^pβ∈Rp from its noisy linear observations y=Xβ+wy= X\beta+ wy=Xβ+w, where w∼N(0,σw2In×n)w \sim N(0, \sigma_w^2 I_{n\times n})w∼N(0,σw2​In×n​), under the following high-dimensional asymptotic regime: given a fixed number δ\deltaδ, p→∞p \rightarrow \inftyp→∞, while n/p→δn/p \rightarrow \deltan/p→δ. We consider the popular class of ℓq\ell_qℓq​-regularized least squares (LQLS) estimators, a.k.a. bridge, given by the optimization problem: \begin{equation*} \hat{\beta} (\lambda, q ) \in \arg\min_\beta \frac{1}{2} \|y-X\beta\|_2^2+ \lambda \|\beta\|_q^q, \end{equation*} and characterize the almost sure limit of 1p∥β^(λ,q)−β∥22\frac{1}{p} \|\hat{\beta} (\lambda, q )- \beta\|_2^2p1​∥β^​(λ,q)−β∥22​. The expression we derive for this limit does not have explicit forms and hence are not useful in comparing different algorithms, or providing information in evaluating the effect of δ\deltaδ or sparsity level of β\betaβ. To simplify the expressions, researchers have considered the ideal "no-noise" regime and have characterized the values of δ\deltaδ for which the almost sure limit is zero. This is known as the phase transition analysis. In this paper, we first perform the phase transition analysis of LQLS. Our results reveal some of the limitations and misleading features of the phase transition analysis. To overcome these limitations, we propose the study of these algorithms under the low noise regime. Our new analysis framework not only sheds light on the results of the phase transition analysis, but also makes an accurate comparison of different regularizers possible.

View on arXiv
Comments on this paper