mDLAM: Neural Network Training via monotonous Deep Learning Alternating Minimization
This is an implementation of monotonous Deep Learning Alternating Minimization(mDLAM) for the neural network training problem, as described in our paper:
Junxiang Wang, Hongyi Li, and Liang Zhao. Accelerated Gradient-free Neural Network Training by Multi-convex Alternating Optimization. (Neurocomputing 2022)
Requirements
torch==1.8.1
numpy==1.21.2
Run the Demo
python mDLAM.py
Data
Four benchmark datasets Cora, PubMed, Citeseer, and Coauthor-CS are included in this package.
Cite
Please cite our following paper if you use our MLP code in your own work:
@inproceedings{wang2022mdlam,
author = {Wang, Junxiang, Li, Hongyi and Zhao, Liang},
title = {Accelerated Gradient-free Neural Network Training by Multi-convex Alternating Optimization},
year = {2022},
booktitle = {Neurocomputing},
}