Stochastic gradient descent in high dimensions for multi-spiked tensor PCA

• URL: http://arxiv.org/abs/2410.18162v1
• Date: Wed, 23 Oct 2024 17:20:41 GMT
• Title: Stochastic gradient descent in high dimensions for multi-spiked tensor PCA
• Authors: Gérard Ben Arous, Cédric Gerbelot, Vanessa Piccolo,
• Abstract summary: We study the dynamics in high dimensions of online gradient descent for the multi-spiked tensor model. We find that the spikes are recovered sequentially in a process we term "sequential elimination"
• Score: 8.435118770300999
• License:
• Abstract: We study the dynamics in high dimensions of online stochastic gradient descent for the multi-spiked tensor model. This multi-index model arises from the tensor principal component analysis (PCA) problem with multiple spikes, where the goal is to estimate $r$ unknown signal vectors within the $N$-dimensional unit sphere through maximum likelihood estimation from noisy observations of a $p$-tensor. We determine the number of samples and the conditions on the signal-to-noise ratios (SNRs) required to efficiently recover the unknown spikes from natural random initializations. We show that full recovery of all spikes is possible provided a number of sample scaling as $N^{p-2}$, matching the algorithmic threshold identified in the rank-one case [Ben Arous, Gheissari, Jagannath 2020, 2021]. Our results are obtained through a detailed analysis of a low-dimensional system that describes the evolution of the correlations between the estimators and the spikes, while controlling the noise in the dynamics. We find that the spikes are recovered sequentially in a process we term "sequential elimination": once a correlation exceeds a critical threshold, all correlations sharing a row or column index become sufficiently small, allowing the next correlation to grow and become macroscopic. The order in which correlations become macroscopic depends on their initial values and the corresponding SNRs, leading to either exact recovery or recovery of a permutation of the spikes. In the matrix case, when $p=2$, if the SNRs are sufficiently separated, we achieve exact recovery of the spikes, whereas equal SNRs lead to recovery of the subspace spanned by the spikes.

Related papers

• Permutation recovery of spikes in noisy high-dimensional tensor estimation [8.435118770300999]
  We study the dynamics of flow in high dimensions for the multi-spiked tensor problem. Our work builds on our companion paper [Ben A, Gerbelot, Langevin], which determines the sample separation conditions for the SNRs necessary for ensuring exact recovery.
  arXiv  Detail & Related papers  (2024-12-19T08:59:49Z)
• Modeling High-Dimensional Dependent Data in the Presence of Many Explanatory Variables and Weak Signals [0.0]
  This article considers a novel and widely applicable approach to modeling high-dimensional dependent data when a large number of explanatory variables are available and the signal-to-noise ratio is low.
  arXiv  Detail & Related papers  (2024-12-06T02:54:31Z)
• Langevin dynamics for high-dimensional optimization: the case of multi-spiked tensor PCA [8.435118770300999]
  We show that the sample complexity required for recovering the spike associated with the largest SNR matches the well-known algorithmic threshold for the single-spike case. As a key step, we provide a detailed characterization of the spikes and interactions that capture the high-dimensional trajectory dynamics.
  arXiv  Detail & Related papers  (2024-08-12T12:09:25Z)
• Computational-Statistical Gaps in Gaussian Single-Index Models [77.1473134227844]
  Single-Index Models are high-dimensional regression problems with planted structure. We show that computationally efficient algorithms, both within the Statistical Query (SQ) and the Low-Degree Polynomial (LDP) framework, necessarily require $Omega(dkstar/2)$ samples.
  arXiv  Detail & Related papers  (2024-03-08T18:50:19Z)
• Retire: Robust Expectile Regression in High Dimensions [3.9391041278203978]
  Penalized quantile and expectile regression methods offer useful tools to detect heteroscedasticity in high-dimensional data. We propose and study (penalized) robust expectile regression (retire) We show that the proposed procedure can be efficiently solved by a semismooth Newton coordinate descent algorithm.
  arXiv  Detail & Related papers  (2022-12-11T18:03:12Z)
• Optimal Online Generalized Linear Regression with Stochastic Noise and Its Application to Heteroscedastic Bandits [88.6139446295537]
  We study the problem of online generalized linear regression in the setting of a generalized linear model with possibly unbounded additive noise. We provide a sharp analysis of the classical follow-the-regularized-leader (FTRL) algorithm to cope with the label noise. We propose an algorithm based on FTRL to achieve the first variance-aware regret bound.
  arXiv  Detail & Related papers  (2022-02-28T08:25:26Z)
• Nonconvex Stochastic Scaled-Gradient Descent and Generalized Eigenvector Problems [98.34292831923335]
  Motivated by the problem of online correlation analysis, we propose the emphStochastic Scaled-Gradient Descent (SSD) algorithm. We bring these ideas together in an application to online correlation analysis, deriving for the first time an optimal one-time-scale algorithm with an explicit rate of local convergence to normality.
  arXiv  Detail & Related papers  (2021-12-29T18:46:52Z)
• Inverting brain grey matter models with likelihood-free inference: a tool for trustable cytoarchitecture measurements [62.997667081978825]
  characterisation of the brain grey matter cytoarchitecture with quantitative sensitivity to soma density and volume remains an unsolved challenge in dMRI. We propose a new forward model, specifically a new system of equations, requiring a few relatively sparse b-shells. We then apply modern tools from Bayesian analysis known as likelihood-free inference (LFI) to invert our proposed model.
  arXiv  Detail & Related papers  (2021-11-15T09:08:27Z)
• Towards an Understanding of Benign Overfitting in Neural Networks [104.2956323934544]
  Modern machine learning models often employ a huge number of parameters and are typically optimized to have zero training loss. We examine how these benign overfitting phenomena occur in a two-layer neural network setting. We show that it is possible for the two-layer ReLU network interpolator to achieve a near minimax-optimal learning rate.
  arXiv  Detail & Related papers  (2021-06-06T19:08:53Z)
• The Generalized Lasso with Nonlinear Observations and Generative Priors [63.541900026673055]
  We make the assumption of sub-Gaussian measurements, which is satisfied by a wide range of measurement models. We show that our result can be extended to the uniform recovery guarantee under the assumption of a so-called local embedding property.
  arXiv  Detail & Related papers  (2020-06-22T16:43:35Z)
• The Heavy-Tail Phenomenon in SGD [7.366405857677226]
  We show that depending on the structure of the Hessian of the loss at the minimum, the SGD iterates will converge to a emphheavy-tailed stationary distribution. We translate our results into insights about the behavior of SGD in deep learning.
  arXiv  Detail & Related papers  (2020-06-08T16:43:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.