SOTAVerified

Second-order methods

Use second-order statistics to process data.

Papers

Showing 126150 of 181 papers

TitleStatusHype
The Challenges of the Nonlinear Regime for Physics-Informed Neural Networks0
The Many Faces of Exponential Weights in Online Learning0
Unlocking FedNL: Self-Contained Compute-Optimized Implementation0
Utility Maximization for Large-Scale Cell-Free Massive MIMO Downlink0
When Does Preconditioning Help or Hurt Generalization?0
KOALA++: Efficient Kalman-Based Optimization of Neural Networks with Gradient-Covariance Products0
A block coordinate descent optimizer for classification problems exploiting convexity0
Accelerated Projected Gradient Method for the Optimization of Cell-Free Massive MIMO Downlink0
Accelerated Training of Federated Learning via Second-Order Methods0
Accelerating SGD for Distributed Deep-Learning Using Approximated Hessian Matrix0
Accelerating Stochastic Probabilistic Inference0
A Chaos Theory Approach to Understand Neural Network Optimization0
A comparison of second-order methods for deep convolutional neural networks0
Adaptive and Optimal Second-order Optimistic Methods for Minimax Optimization0
A Learn-to-Optimize Approach for Coordinate-Wise Step Sizes for Quasi-Newton Methods0
Adaptive Optimization Algorithms for Machine Learning0
Adaptive Second Order Coresets for Data-efficient Machine Learning0
Adaptive Single-Pass Stochastic Gradient Descent in Input Sparsity Time0
A Decentralized Quasi-Newton Method for Dual Formulations of Consensus Optimization0
A Distributed Optimisation Framework Combining Natural Gradient with Hessian-Free for Discriminative Sequence Training0
A Distributed Second-Order Algorithm You Can Trust0
Adversarial Vulnerability as a Consequence of On-Manifold Inseparibility0
A Flexible Tensor Block Coordinate Ascent Scheme for Hypergraph Matching0
A Generic Approach for Escaping Saddle points0
Gram-Gauss-Newton Method: Learning Overparameterized Neural Networks for Regression Problems0
Show:102550
← PrevPage 6 of 8Next →

No leaderboard results yet.