SOTAVerified

Second-order methods

Use second-order statistics to process data.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 101–125 of 181 papers

TitleStatusHype
A Distributed Optimisation Framework Combining Natural Gradient with Hessian-Free for Discriminative Sequence Training—0
Learning-Augmented Sketches for Hessians—0
Distributed Second Order Methods with Fast Rates and Compressed Communication—0
Kronecker-factored Quasi-Newton Methods for Deep Learning—0
Adaptive Single-Pass Stochastic Gradient Descent in Input Sparsity Time—0
A Chaos Theory Approach to Understand Neural Network Optimization—0
Noise and Fluctuation of Finite Learning Rate Stochastic Gradient Descent—0
Utility Maximization for Large-Scale Cell-Free Massive MIMO Downlink—0
Second-order Neural Network Training Using Complex-step Directional Derivative—0
Debiasing Distributed Second Order Optimization with Surrogate Sketching and Scaled Regularization—0
Second-Order Information in Non-Convex Stochastic Optimization: Power and Limitations—0
When Does Preconditioning Help or Hurt Generalization?—0
A block coordinate descent optimizer for classification problems exploiting convexity—0
Enhance Curvature Information by Structured Stochastic Quasi-Newton Methods—0
Learning Rates as a Function of Batch Size: A Random Matrix Theory Approach to Neural Network Training—0
Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian OptimizationCode2
SONIA: A Symmetric Blockwise Truncated Optimization Algorithm—0
Asymptotic Analysis of Conditioned Stochastic Gradient DescentCode0
On the Promise of the Stochastic Generalized Gauss-Newton Method for Training DNNsCode0
ADAHESSIAN: An Adaptive Second Order Optimizer for Machine LearningCode1
MiLeNAS: Efficient Neural Architecture Search via Mixed-Level ReformulationCode1
Critical Point-Finding Methods Reveal Gradient-Flat Regions of Deep Network Losses—0
Stochastic Subspace Cubic Newton Method—0
DDPNOpt: Differential Dynamic Programming Neural Optimizer—0
Low Rank Saddle Free Newton: A Scalable Method for Stochastic Nonconvex OptimizationCode1
Show:102550
← PrevPage 5 of 8Next →

No leaderboard results yet.