Resultats de la cerca - Stich, Sebastian

On low complexity Acceleration Techniques for Randomized Optimization: Supplementary Online Material per Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

On Communication Compression for Distributed Optimization on Heterogeneous Data per Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

On the Convergence of SGD with Biased Gradients per Ajalloeian, Ahmad, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Tackling benign nonconvexity with smoothing and stochastic gradients per Vardhan, Harsh, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Adaptive SGD with Polyak stepsize and Line-search: Robust Convergence and Variance Reduction per Jiang, Xiaowen, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Scalable Decentralized Learning with Teleportation per Takezawa, Yuki, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Sequential Subspace Noise Injection Prevents Accuracy Collapse in Certified Unlearning per Dolgova, Polina, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

On the Stability of Nonlinear Dynamics in GD and SGD: Beyond Quadratic Potentials per Mulayoff, Rotem, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Forgetting Has Neighbors: Localized Collateral Forgetting in Machine Unlearning per Dolgova, Polina, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

The Error-Feedback Framework: Better Rates for SGD with Delayed Gradients and Compressed Communication per Stich, Sebastian U., Karimireddy, Sai Praneeth

Obtenir text complet
Obtenir text complet

Characterizing & Finding Good Data Orderings for Fast Convergence of Sequential Gradient Methods per Mohtashami, Amirkeivan, Stich, Sebastian, Jaggi, Martin

Obtenir text complet
Obtenir text complet

Special Properties of Gradient Descent with Large Learning Rates per Mohtashami, Amirkeivan, Jaggi, Martin, Stich, Sebastian

Obtenir text complet
Obtenir text complet

EControl: Fast Distributed Optimization with Compression and Error Control per Gao, Yuan, Islamov, Rustem, Stich, Sebastian

Obtenir text complet
Obtenir text complet

Critical Parameters for Scalable Distributed Learning with Large Batches and Asynchronous Updates per Stich, Sebastian U., Mohtashami, Amirkeivan, Jaggi, Martin

Obtenir text complet
Obtenir text complet

Masked Training of Neural Networks with Partial Gradients per Mohtashami, Amirkeivan, Jaggi, Martin, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

An Improved Analysis of Gradient Tracking for Decentralized Machine Learning per Koloskova, Anastasia, Lin, Tao, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Sharper Convergence Guarantees for Asynchronous SGD for Distributed and Federated Learning per Koloskova, Anastasia, Stich, Sebastian U., Jaggi, Martin

Obtenir text complet
Obtenir text complet

Revisiting Gradient Clipping: Stochastic bias and tight convergence guarantees per Koloskova, Anastasia, Hendrikx, Hadrien, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Locally Adaptive Federated Learning per Mukherjee, Sohom, Loizou, Nicolas, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet

Non-convex Stochastic Composite Optimization with Polyak Momentum per Gao, Yuan, Rodomanov, Anton, Stich, Sebastian U.

Obtenir text complet
Obtenir text complet