English

Principled Curriculum Learning using Parameter Continuation Methods

Machine Learning 2025-07-31 v1 Artificial Intelligence

Abstract

In this work, we propose a parameter continuation method for the optimization of neural networks. There is a close connection between parameter continuation, homotopies, and curriculum learning. The methods we propose here are theoretically justified and practically effective for several problems in deep neural networks. In particular, we demonstrate better generalization performance than state-of-the-art optimization techniques such as ADAM for supervised and unsupervised learning tasks.

Keywords

Cite

@article{arxiv.2507.22089,
  title  = {Principled Curriculum Learning using Parameter Continuation Methods},
  author = {Harsh Nilesh Pathak and Randy Paffenroth},
  journal= {arXiv preprint arXiv:2507.22089},
  year   = {2025}
}
R2 v1 2026-07-01T04:24:37.230Z