[2604.02393] Dynamical structure of vanishing gradient and overfitting in multi-layer perceptrons

Computer Science > Machine Learning arXiv:2604.02393 (cs) [Submitted on 2 Apr 2026] Title:Dynamical structure of vanishing gradient and overfitting in multi-layer perceptrons Authors:Alex Alì Maleknia, Yuzuru Sato View a PDF of the paper titled Dynamical structure of vanishing gradient and overfitting in multi-layer perceptrons, by Alex Al\`i Maleknia and Yuzuru Sato View PDF HTML (experimental) Abstract:Vanishing gradient and overfitting are two of the most extensively studied problems in the literature about machine learning. However, they are frequently considered in some asymptotic setting, which obscure the underlying dynamical mechanisms responsible for their emergence. In this paper, we aim to provide a clear dynamical description of learning in multi-layer perceptrons. To this end, we introduce a minimal model, inspired by studies by Fukumizu and Amari, to investigate vanishing gradients and overfitting in MLPs trained via gradient descent. Within this model, we show that the learning dynamics may pass through plateau regions and near-optimal regions during training, both of which consist of saddle structures, before ultimately converging to the overfitting region. Under suitable conditions on the training dataset, we prove that, with high probability, the overfitting region collapses to a single attractor modulo symmetry, which corresponds to the overfitting. Moreover, we show that any MLP trained on a finite noisy dataset cannot converge to the theoretical optimu...