Mher Safaryan
11 papers · 2021–2025 · 5 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓+1 more ↓ Show less ↑
🌍 Conference Polyglot · Silver 🌉 Interdisciplinary Bridge · Bronze 🗺️ Taxonomy Completionist (18) 💎 Century Club · Bronze 🔥 Unstoppable · Bronze
🗃️
Keyword Collector
(53)
Top co-authors
Research topics
Machine Learning
(11)
Optimization & Theory · Machine Learning
(11)
Mathematics & Optimization
(7)
Optimization
(7)
Application Areas
(5)
Artificial Intelligence
(2)
Deep Learning
(2)
Learning Paradigms · Machine Learning
(2)
Optimization & Theory · Deep Learning
(2)
Core AI
(1)
Learning Paradigms · Artificial Intelligence
(1)
Learning Types
(1)
Keywords
distributed optimization
(5)
model compression
(3)
stochastic gradient descent
(3)
communication compression
(3)
federated learning
(2)
variance reduction
(2)
second-order optimization
(2)
convergence analysis
(2)
adaptive optimization
(2)
distributed learning
(2)
gradient compression
(2)
communication efficiency
(2)
error feedback
(2)
non-convex optimization
(2)
memory efficiency
(2)
distributed training
(1)
neural network optimization
(1)
sparse recovery
(1)
convergence guarantee
(1)
linear convergence
(1)
Papers
LDAdam: Adaptive Optimization from Low-Dimensional Gradient Statistics
ICLR 2025
AsGrad: A Sharp Unified Analysis of Asynchronous-SGD Algorithms
AISTATS 2024
MicroAdam: Accurate Adaptive Optimization with Low Space Overhead and Provable Convergence
NeurIPS 2024
The Iterative Optimal Brain Surgeon: Faster Sparse Recovery by Leveraging Second-Order Information
NeurIPS 2024
Knowledge Distillation Performs Partial Variance Reduction
NeurIPS 2023
On Biased Compression for Distributed Learning
JMLR 2023
Basis Matters: Better Communication-Efficient Second Order Methods for Federated Learning
AISTATS 2022
FedNL: Making Newton-Type Methods Applicable to Federated Learning
ICML 2022
Theoretically Better and Numerically Faster Distributed Optimization with Smoothness-Aware Quantization Techniques
NeurIPS 2022
Smoothness Matrices Beat Smoothness Constants: Better Communication Compression Techniques for Distributed Optimization
NeurIPS 2021
Stochastic Sign Descent Methods: New Algorithms and Better Theory
ICML 2021