Large-Scale Riemannian Meta-Optimization via Subspace Adaptation

• URL: http://arxiv.org/abs/2501.15235v2
• Date: Thu, 06 Feb 2025 01:35:55 GMT
• Title: Large-Scale Riemannian Meta-Optimization via Subspace Adaptation
• Authors: Peilin Yu, Yuwei Wu, Zhi Gao, Xiaomeng Fan, Yunde Jia,
• Abstract summary: We propose an efficient method that significantly reduces the memory burden for large-scale optimization. Our method reduces the model memory consumption by six orders of magnitude when optimizing an mainstream deep neural network.
• Score: 39.75524650528829
• License:
• Abstract: Riemannian meta-optimization provides a promising approach to solving non-linear constrained optimization problems, which trains neural networks as optimizers to perform optimization on Riemannian manifolds. However, existing Riemannian meta-optimization methods take up huge memory footprints in large-scale optimization settings, as the learned optimizer can only adapt gradients of a fixed size and thus cannot be shared across different Riemannian parameters. In this paper, we propose an efficient Riemannian meta-optimization method that significantly reduces the memory burden for large-scale optimization via a subspace adaptation scheme. Our method trains neural networks to individually adapt the row and column subspaces of Riemannian gradients, instead of directly adapting the full gradient matrices in existing Riemannian meta-optimization methods. In this case, our learned optimizer can be shared across Riemannian parameters with different sizes. Our method reduces the model memory consumption by six orders of magnitude when optimizing an orthogonal mainstream deep neural network (e.g., ResNet50). Experiments on multiple Riemannian tasks show that our method can not only reduce the memory consumption but also improve the performance of Riemannian meta-optimization.

Related papers

• Riemannian Bilevel Optimization [35.42472057648458]
  We focus in particular on batch and gradient-based methods, with the explicit goal of avoiding second-order information. We propose and analyze $mathrmRF2SA$, a method that leverages first-order gradient information. We provide explicit convergence rates for reaching $epsilon$-stationary points under various setups.
  arXiv  Detail & Related papers  (2024-05-22T20:49:01Z)
• Streamlining in the Riemannian Realm: Efficient Riemannian Optimization with Loopless Variance Reduction [4.578425862931332]
  This study focuses on the crucial reduction mechanism used in both Euclidean and Riemannian settings. Motivated by Euclidean methods, we introduce R-based methods to replace the outer loop with computation triggered by a coin flip. Using R- as a framework, we demonstrate its applicability to various important settings.
  arXiv  Detail & Related papers  (2024-03-11T12:49:37Z)
• FORML: A Riemannian Hessian-free Method for Meta-learning on Stiefel Manifolds [4.757859522106933]
  This paper introduces a Hessian-free approach that uses a first-order approximation of derivatives on the Stiefel manifold. Our method significantly reduces the computational load and memory footprint.
  arXiv  Detail & Related papers  (2024-02-28T10:57:30Z)
• Riemannian Optimization for Variance Estimation in Linear Mixed Models [0.0]
  We take a completely novel view on parameter estimation in linear mixed models by exploiting the intrinsic geometry of the parameter space. Our approach yields a higher quality of the variance parameter estimates compared to existing approaches.
  arXiv  Detail & Related papers  (2022-12-18T13:08:45Z)
• On Riemannian Approach for Constrained Optimization Model in Extreme Classification Problems [2.7436792484073638]
  A constrained optimization problem is formulated as an optimization problem on matrix manifold. The proposed approach is tested on several real world large scale multi-label datasets.
  arXiv  Detail & Related papers  (2021-09-30T11:28:35Z)
• Implicit differentiation for fast hyperparameter selection in non-smooth convex learning [87.60600646105696]
  We study first-order methods when the inner optimization problem is convex but non-smooth. We show that the forward-mode differentiation of proximal gradient descent and proximal coordinate descent yield sequences of Jacobians converging toward the exact Jacobian.
  arXiv  Detail & Related papers  (2021-05-04T17:31:28Z)
• Unified Convergence Analysis for Adaptive Optimization with Moving Average Estimator [75.05106948314956]
  We show that an increasing large momentum parameter for the first-order moment is sufficient for adaptive scaling. We also give insights for increasing the momentum in a stagewise manner in accordance with stagewise decreasing step size.
  arXiv  Detail & Related papers  (2021-04-30T08:50:24Z)
• Automatic differentiation for Riemannian optimization on low-rank matrix and tensor-train manifolds [71.94111815357064]
  In scientific computing and machine learning applications, matrices and more general multidimensional arrays (tensors) can often be approximated with the help of low-rank decompositions. One of the popular tools for finding the low-rank approximations is to use the Riemannian optimization.
  arXiv  Detail & Related papers  (2021-03-27T19:56:00Z)
• Global Optimization of Gaussian processes [52.77024349608834]
  We propose a reduced-space formulation with trained Gaussian processes trained on few data points. The approach also leads to significantly smaller and computationally cheaper sub solver for lower bounding. In total, we reduce time convergence by orders of orders of the proposed method.
  arXiv  Detail & Related papers  (2020-05-21T20:59:11Z)
• Learning to Guide Random Search [111.71167792453473]
  We consider derivative-free optimization of a high-dimensional function that lies on a latent low-dimensional manifold. We develop an online learning approach that learns this manifold while performing the optimization. We empirically evaluate the method on continuous optimization benchmarks and high-dimensional continuous control problems.
  arXiv  Detail & Related papers  (2020-04-25T19:21:14Z)

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.