Local minima of the empirical risk in high dimension: General theorems and convex examples

• URL: http://arxiv.org/abs/2502.01953v1
• Date: Tue, 04 Feb 2025 03:02:24 GMT
• Title: Local minima of the empirical risk in high dimension: General theorems and convex examples
• Authors: Kiana Asgari, Andrea Montanari, Basil Saeed,
• Abstract summary: We consider a general model for high-dimensional empirical risk whereby the data vectors $mathbfxi$ are $d- minimization. We derive sharps on the estimation and prediction error.
• Score: 8.748904058015574
• License:
• Abstract: We consider a general model for high-dimensional empirical risk minimization whereby the data $\mathbf{x}_i$ are $d$-dimensional isotropic Gaussian vectors, the model is parametrized by $\mathbf{\Theta}\in\mathbb{R}^{d\times k}$, and the loss depends on the data via the projection $\mathbf{\Theta}^\mathsf{T}\mathbf{x}_i$. This setting covers as special cases classical statistics methods (e.g. multinomial regression and other generalized linear models), but also two-layer fully connected neural networks with $k$ hidden neurons. We use the Kac-Rice formula from Gaussian process theory to derive a bound on the expected number of local minima of this empirical risk, under the proportional asymptotics in which $n,d\to\infty$, with $n\asymp d$. Via Markov's inequality, this bound allows to determine the positions of these minimizers (with exponential deviation bounds) and hence derive sharp asymptotics on the estimation and prediction error. In this paper, we apply our characterization to convex losses, where high-dimensional asymptotics were not (in general) rigorously established for $k\ge 2$. We show that our approach is tight and allows to prove previously conjectured results. In addition, we characterize the spectrum of the Hessian at the minimizer. A companion paper applies our general result to non-convex examples.

Related papers

• Which exceptional low-dimensional projections of a Gaussian point cloud can be found in polynomial time? [8.74634652691576]
  We study the subset $mathscrF_m,alpha$ of distributions that can be realized by a class of iterative algorithms. Non-rigorous methods from statistical physics yield an indirect characterization of $mathscrF_m,alpha$ in terms of a generalized Parisi formula.
  arXiv  Detail & Related papers  (2024-06-05T05:54:56Z)
• 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)
• A Unified Framework for Uniform Signal Recovery in Nonlinear Generative Compressed Sensing [68.80803866919123]
  Under nonlinear measurements, most prior results are non-uniform, i.e., they hold with high probability for a fixed $mathbfx*$ rather than for all $mathbfx*$ simultaneously. Our framework accommodates GCS with 1-bit/uniformly quantized observations and single index models as canonical examples. We also develop a concentration inequality that produces tighter bounds for product processes whose index sets have low metric entropy.
  arXiv  Detail & Related papers  (2023-09-25T17:54:19Z)
• Effective Minkowski Dimension of Deep Nonparametric Regression: Function Approximation and Statistical Theories [70.90012822736988]
  Existing theories on deep nonparametric regression have shown that when the input data lie on a low-dimensional manifold, deep neural networks can adapt to intrinsic data structures. This paper introduces a relaxed assumption that input data are concentrated around a subset of $mathbbRd$ denoted by $mathcalS$, and the intrinsic dimension $mathcalS$ can be characterized by a new complexity notation -- effective Minkowski dimension.
  arXiv  Detail & Related papers  (2023-06-26T17:13:31Z)
• Universality laws for Gaussian mixtures in generalized linear models [22.154969876570238]
  We investigate the joint statistics of the family of generalized linear estimators $(Theta_1, dots, Theta_M)$. This allow us to prove the universality of different quantities of interest, such as the training and generalization errors. We discuss the applications of our results to different machine learning tasks of interest, such as ensembling and uncertainty.
  arXiv  Detail & Related papers  (2023-02-17T15:16:06Z)
• Dimension free ridge regression [10.434481202633458]
  We revisit ridge regression on i.i.d. data in terms of the bias and variance of ridge regression in terms of the bias and variance of an equivalent' sequence model. As a new application, we obtain a completely explicit and sharp characterization of ridge regression for Hilbert covariates with regularly varying spectrum.
  arXiv  Detail & Related papers  (2022-10-16T16:01:05Z)
• $p$-Generalized Probit Regression and Scalable Maximum Likelihood Estimation via Sketching and Coresets [74.37849422071206]
  We study the $p$-generalized probit regression model, which is a generalized linear model for binary responses. We show how the maximum likelihood estimator for $p$-generalized probit regression can be approximated efficiently up to a factor of $(1+varepsilon)$ on large data.
  arXiv  Detail & Related papers  (2022-03-25T10:54:41Z)
• Minimax Optimal Quantization of Linear Models: Information-Theoretic Limits and Efficient Algorithms [59.724977092582535]
  We consider the problem of quantizing a linear model learned from measurements. We derive an information-theoretic lower bound for the minimax risk under this setting. We show that our method and upper-bounds can be extended for two-layer ReLU neural networks.
  arXiv  Detail & Related papers  (2022-02-23T02:39:04Z)
• Universality of empirical risk minimization [12.764655736673749]
  Consider supervised learning from i.i.d. samples where $boldsymbol x_i inmathbbRp$ are feature vectors and $y in mathbbR$ are labels. We study empirical risk universality over a class of functions that are parameterized by $mathsfk.
  arXiv  Detail & Related papers  (2022-02-17T18:53:45Z)
• A Statistical Learning View of Simple Kriging [0.0]
  We analyze the simple Kriging task from a statistical learning perspective. The goal is to predict the unknown values it takes at any other location with minimum quadratic risk. We prove non-asymptotic bounds of order $O_mathbbP (1/sqrtn)$ for the excess risk of a plug-in predictive rule mimicking the true minimizer.
  arXiv  Detail & Related papers  (2022-02-15T12:46:43Z)
• Agnostic Learning of a Single Neuron with Gradient Descent [92.7662890047311]
  We consider the problem of learning the best-fitting single neuron as measured by the expected square loss. For the ReLU activation, our population risk guarantee is $O(mathsfOPT1/2)+epsilon$. For the ReLU activation, our population risk guarantee is $O(mathsfOPT1/2)+epsilon$.
  arXiv  Detail & Related papers  (2020-05-29T07:20:35Z)

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.