ASCEND
BY NTHRYS

NTHRYS › PhD Assistance › Optimization Science

Optimization Science

Field
▾
Category
▾

Optimization Science

Select a category to explore research frontiers

Loading categories...

Research Frontiers in Stochastic Gradient Descent Convergence Analysis

Theoretical investigation of convergence rates and variance reduction techniques for stochastic gradient methods in large-scale learning.

Noise-Induced Acceleration in Non-Convex Landscapes
Implicit Regularization Through Stochastic Gradient Geometry
Critical Point Escape: SGD's Dance with Saddle Dynamics
Variance Reduction at the Edge of Chaos
Temporal Correlations in Mini-Batch Sampling Effects
Oscillatory Convergence in Heterogeneous Federated Learning
Adaptive Curvature Sensing in Gradient Noise Fields
Phase Transitions in Loss Surface Exploration
Stochastic Resonance in Deep Network Training Trajectories
Gradient Noise Anisotropy and Basin Selection Mechanisms

All Optimization Science PhD categories