ASCEND
BY NTHRYS

NTHRYS › PhD Assistance › Mathematical Optimization

Mathematical Optimization

Field
▾
Category
▾

Mathematical Optimization

Select a category to explore research frontiers

Loading categories...

Research Frontiers in Stochastic Gradient Descent Methods

Analysis and development of variance-reduced and adaptive SGD variants for large-scale machine learning optimization.

Adaptive Momentum Landscapes in Non-Convex Optimization
Variance Reduction Under Extreme Sparsity Regimes
Implicit Regularization Through Gradient Noise Geometry
Second-Order Information Recovery from Stochastic Trajectories
Decentralized SGD in Heterogeneous Network Topologies
Escaping Saddle Points with Structured Noise Injection
Momentum Acceleration Beyond Smooth Convexity Assumptions
Federated Learning Under Byzantine Gradient Corruption
Curvature Estimation From Single-Pass Stochastic Samples
Memory-Efficient Preconditioning for Ultra-High-Dimensional Problems

All Mathematical Optimization PhD categories