Temporal Organization of Weight Corrections during Adam Optimization of Neural Networks: Dataset

This dataset accompanies the research study "Temporal Organization of Weight Corrections during Adam Optimization of Neural Networks". The dataset contains experimental results obtained during the analysis of temporal organization of neural network weight corrections under Adam optimization. The data include temporal trajectories of normalized weight corrections, spatial synchronization observables, cross-correlation analysis results, and auxiliary experimental parameters for different neural network configurations. Experiments were performed using CIFAR-10 datasets and multilayer neural network architectures with different input dimensions, optimizer parameters, and training conditions. The dataset is provided to support reproducibility and further investigation of dynamic regime mapping, temporal synchronization phenomena, and nonlinear properties of neural network optimization dynamics.