Study Establishes Sharp Lower Bounds for Averaged SGD Inference Complexity
A new paper establishes sharp lower bounds for Gaussian approximation of Polyak-Ruppert averaged stochastic gradient descent with polynomially decreasing step sizes.
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-11-03 05:00 · arXiv stat.ML
Lower Bounds on Inference Complexity for Polyak-Ruppert Averaged Stochastic Gradient Descent