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. 2006.02601
16
39

On the Minimax Optimality of the EM Algorithm for Learning Two-Component Mixed Linear Regression

4 June 2020
Jeongyeol Kwon
Nhat Ho
Constantine Caramanis
ArXivPDFHTML
Abstract

We study the convergence rates of the EM algorithm for learning two-component mixed linear regression under all regimes of signal-to-noise ratio (SNR). We resolve a long-standing question that many recent results have attempted to tackle: we completely characterize the convergence behavior of EM, and show that the EM algorithm achieves minimax optimal sample complexity under all SNR regimes. In particular, when the SNR is sufficiently large, the EM updates converge to the true parameter θ∗\theta^{*}θ∗ at the standard parametric convergence rate O((d/n)1/2)\mathcal{O}((d/n)^{1/2})O((d/n)1/2) after O(log⁡(n/d))\mathcal{O}(\log(n/d))O(log(n/d)) iterations. In the regime where the SNR is above O((d/n)1/4)\mathcal{O}((d/n)^{1/4})O((d/n)1/4) and below some constant, the EM iterates converge to a O(SNR−1(d/n)1/2)\mathcal{O}({\rm SNR}^{-1} (d/n)^{1/2})O(SNR−1(d/n)1/2) neighborhood of the true parameter, when the number of iterations is of the order O(SNR−2log⁡(n/d))\mathcal{O}({\rm SNR}^{-2} \log(n/d))O(SNR−2log(n/d)). In the low SNR regime where the SNR is below O((d/n)1/4)\mathcal{O}((d/n)^{1/4})O((d/n)1/4), we show that EM converges to a O((d/n)1/4)\mathcal{O}((d/n)^{1/4})O((d/n)1/4) neighborhood of the true parameters, after O((n/d)1/2)\mathcal{O}((n/d)^{1/2})O((n/d)1/2) iterations. Notably, these results are achieved under mild conditions of either random initialization or an efficiently computable local initialization. By providing tight convergence guarantees of the EM algorithm in middle-to-low SNR regimes, we fill the remaining gap in the literature, and significantly, reveal that in low SNR, EM changes rate, matching the n−1/4n^{-1/4}n−1/4 rate of the MLE, a behavior that previous work had been unable to show.

View on arXiv
Comments on this paper