Publications & Preprints

SILAGE: Memory-Efficient, Full-Gradient-Free Nonconvex Optimization for Nested Finite Sums
Igor Sokolov, Laurent Condat, Peter Richtárik
arXiv
Jun 2026
PDF Code arXiv
Rennala-NSGD: Asynchronous Stochastic Optimization Beyond Euclidean Geometry
Igor Sokolov, Alexander Tyurin, Peter Richtárik
Under review
Mar 2026
Improved Convergence in Parameter-Agnostic Error Feedback through Momentum
Abdurakhmon Sadiev, Yury Demidovich, Igor Sokolov, Grigory Malinovsky, Sarit Khirirat, Peter Richtárik
arXiv
Nov 2025
PDF Cite Code arXiv
Bernoulli-LoRA: A Theoretical Framework for Randomized Low-Rank Adaptation
Igor Sokolov, Abdurakhmon Sadiev, Yury Demidovich, Fawaz S Al-Qahtani, Peter Richtárik
arXiv
Aug 2025
PDF Cite arXiv
MARINA-P: Superior Performance in Non-smooth Federated Optimization with Adaptive Stepsizes
Igor Sokolov, Peter Richtárik
arXiv
Dec 2024
PDF Cite Code arXiv
Cohort Squeeze: Beyond a Single Communication Round per Cohort in Cross-Device Federated Learning
Kai Yi, Timur Kharisov, Igor Sokolov, Peter Richtárik
FL@FM-NeurIPS 2024 Workshop
Jun 2024
PDF Cite arXiv FL@FM-NeurIPS 2024
A Guide Through the Zoo of Biased SGD
Yury Demidovich, Grigory Malinovsky, Igor Sokolov, Peter Richtárik
NeurIPS 2023
May 2023
PDF Cite Code Poster NeurIPS 23 arXiv
Don't Compress Gradients in Random Reshuffling: Compress Gradient Differences
Abdurakhmon Sadiev, Grigory Malinovsky, Eduard Gorbunov, Igor Sokolov, Ahmed Khaled, Konstantin Pavlovich Burlachenko, Peter Richtárik
NeurIPS 2024
Jun 2022
PDF Cite Code Poster NeurIPS 24 arXiv FL-ICML 23
3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation
Peter Richtárik, Igor Sokolov, Elnur Gasanov, Ilyas Fatkhullin, Zhize Li, Eduard Gorbunov
ICML 2022
Feb 2022
PDF Cite Code Poster ICML 22 arXiv
EF21 with Bells & Whistles: Practical Algorithmic Extensions of Modern Error Feedback
Ilyas Fatkhullin, Igor Sokolov, Eduard Gorbunov, Zhize Li, Peter Richtárik
JMLR 2025
Oct 2021
PDF Cite Code Poster JMLR 25 arXiv OPT-ML 21 FLOW video FLOW slides
EF21: A New, Simpler, Theoretically Better, and Practically Faster Error Feedback
Peter Richtárik, Igor Sokolov, Ilyas Fatkhullin
NeurIPS 2021
Jun 2021
PDF Cite Code Poster NeurIPS 21 arXiv FL-ICML 21 FL-ICML 21 slides FLOW video FLOW slides