tlooto
Sign up
Stochastic Gradient Optimization TechniquesAdvanced Neural Network ApplicationsAdvanced Data Compression Techniques

Adaptive µ-law Gradient Quantization for Training MLPs and CNNs

M. Dubljanin, S. Panić, M. Savić, S. Milosavljevic

2026.1.1ADVANCES IN ELECTRICAL AND COMPUTER ENGINEERING

DOI: 10.4316/aece.2026.01006

Abstract

This paper presents an adaptive µ-quasilogarithmic gradient quantization framework aimed at reducing memory and computational demands during deep neural network training. The appr ...

Citation format

DUBLJANIN, M., et al. Adaptive µ-law gradient quantization for training MLPs and CNNs. ADVANCES IN ELECTRICAL AND COMPUTER ENGINEERING, 2026, 26(1): 55–64.