A Unified Approach to Controlling Implicit Regularization Using Mirror Descent

A Unified Approach to Controlling Implicit Regularization Using Mirror Descent

Inspired by the remarkable performance of deep neural networks, understanding the generalization performance of overparameterized models and the effect of optimization algorithms on it has become an increasingly popular question. In particular, there has been substantial effort to characterize the s...

書目詳細資料
主要作者:	Sun, Haoyuan
其他作者:	Jadbabaie, Ali
格式:	Thesis
出版:	Massachusetts Institute of Technology 2023
在線閱讀:	https://hdl.handle.net/1721.1/151464 https://orcid.org/0000-0002-6203-0198

相似書籍

Implicit regularization in matrix sensing via mirror descent
由: Wu, F, et al.
出版: (2021)

Learning mirror maps in policy mirror descent
由: Alfano, C, et al.
出版: (2025)

A continuous-time mirror descent approach to sparse phase retrieval
由: Wu, F, et al.
出版: (2020)

The statistical complexity of early-stopped mirror descent
由: Kanade, V, et al.
出版: (2023)

The statistical complexity of early-stopped mirror descent
由: Vaškevičius, T, et al.
出版: (2020)

Mirror descent for high-dimensional statistical models
由: Wu, F
出版: (2021)

Mirror Descent view for Neural Network quantization
由: Ajanthan, T, et al.
出版: (2021)

Distributed Online Optimization in Dynamic Environments Using Mirror Descent
由: Shahrampour, Shahin
出版: (2018)

Iterative Regularization via Dual Diagonal Descent
由: Garrigos, Guillaume, et al.
出版: (2018)

Graph-dependent implicit regularisation for distributed stochastic subgradient descent
由: Richards, D, et al.
出版: (2020)

The implicit adaptation to temporal regularities
由: Shalev, N, et al.
出版: (2017)

Implicit dynamic regularization in deep networks
由: Poggio, Tomaso, et al.
出版: (2020)

Implicit regularization for optimal sparse recovery
由: Vaškevičius, T, et al.
出版: (2019)

A novel framework for policy mirror descent with general parameterization and linear convergence
由: Alfano, C, et al.
出版: (2024)

Representational drift as a result of implicit regularization
由: Aviv Ratzon, et al.
出版: (2024-05-01)

Implicit Regularization and Momentum Algorithms in Nonlinearly Parameterized Adaptive Control and Prediction
由: Boffi, Nicholas M, et al.
出版: (2022)

Optimal convergence rate for exact policy mirror descent in discounted Markov decision processes
由: Johnson, E, et al.
出版: (2023)

A unified framework for consistency of regularized loss minimizers
由: Honorio, Jean, et al.
出版: (2015)

On the Complexity of Steepest Descent, Newton's and Regularized Newton's Methods for Nonconvex Unconstrained Optimization Problems.
由: Cartis, C, et al.
出版: (2010)

Implicit Uniform Rational Spline Reconstruction Method With Regularization Term
由: Shenghan Huang
出版: (2024-01-01)

Nearly minimax-optimal rates for noisy sparse phase retrieval via early-stopped mirror descent
由: Wu, F, et al.
出版: (2022)

The descent of testis and reason for failed descent
由: F Cahit Tanyel
出版: (2004-01-01)

Unified regularity measures for sample-wise learning and generalization
由: Chi Zhang, et al.
出版: (2024-12-01)

A Unified Framework for Regularization Networks and Support Vector Machines
由: Evgeniou, Theodoros, et al.
出版: (2004)

On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
由: Dollar, H, et al.
出版: (2005)

Implicit-factorization preconditioning and iterative solvers for regularized saddle-point systems
由: Dollar, H, et al.
出版: (2006)

Preserved learning of implicit regularities with predictive social cues in older adults
由: Claudia Salera, et al.
出版: (2024-12-01)

Unified formulation of leapfrog complying divergence implicit FDTD method for lossy media
由: Liu, Shuo, et al.
出版: (2024)

Visual control of action in step descent.
由: Cowie, D, et al.
出版: (2008)

Predictive control : a unified approach/
由: 310310 Soeterboek, A. Ronald
出版: (1992)

A machine learning approach for the prediction of top of descent
由: Tan, Benjamin Zhi Yong, et al.
出版: (2021)

When cyclic coordinate descent outperforms randomized coordinate descent
由: Gurbuzbalaban, Mert, et al.
出版: (2019)

Complexity control by gradient descent in deep networks
由: Poggio, Tomaso, et al.
出版: (2021)

Complexity control by gradient descent in deep networks
由: Poggio, Tomaso A, et al.
出版: (2022)

Controlled gradient descent: A control theoretical perspective for optimization
由: Revati Gunjal, et al.
出版: (2024-06-01)

Convergence of Implicit Iterative Processes for Semigroups of Nonlinear Operators Acting in Regular Modular Spaces
由: Wojciech M. Kozlowski
出版: (2024-12-01)

Unified leapfrog CDI-FDTD methods with weakly conditional stability and hybrid implicit-explicit schemes
由: Liu, Shuo, et al.
出版: (2024)

Benchmarking of implicit numerical integration methods for stiff unified constitutive equations in metal forming applications
由: James Dear, et al.
出版: (2024-12-01)

Digital control and estimation : a unified approach /
由: 247164 Middleton, Richard H., et al.
出版: (1990)

The field descent method
由: Bernhard, Schmidt., et al.
出版: (2009)