Enregistré dans:
Détails bibliographiques
Auteurs principaux: Tan, Charlie B., García-Redondo, Inés, Wang, Qiquan, Bronstein, Michael M., Monod, Anthea
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2406.02234
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916464053714944
author Tan, Charlie B.
García-Redondo, Inés
Wang, Qiquan
Bronstein, Michael M.
Monod, Anthea
author_facet Tan, Charlie B.
García-Redondo, Inés
Wang, Qiquan
Bronstein, Michael M.
Monod, Anthea
contents Bounding and predicting the generalization gap of overparameterized neural networks remains a central open problem in theoretical machine learning. There is a recent and growing body of literature that proposes the framework of fractals to model optimization trajectories of neural networks, motivating generalization bounds and measures based on the fractal dimension of the trajectory. Notably, the persistent homology dimension has been proposed to correlate with the generalization gap. This paper performs an empirical evaluation of these persistent homology-based generalization measures, with an in-depth statistical analysis. Our study reveals confounding effects in the observed correlation between generalization and topological measures due to the variation of hyperparameters. We also observe that fractal dimension fails to predict generalization of models trained from poor initializations. We lastly reveal the intriguing manifestation of model-wise double descent in these topological generalization measures. Our work forms a basis for a deeper investigation of the causal relationships between fractal geometry, topological data analysis, and neural network optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2406_02234
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On the Limitations of Fractal Dimension as a Measure of Generalization
Tan, Charlie B.
García-Redondo, Inés
Wang, Qiquan
Bronstein, Michael M.
Monod, Anthea
Machine Learning
Artificial Intelligence
Dynamical Systems
Bounding and predicting the generalization gap of overparameterized neural networks remains a central open problem in theoretical machine learning. There is a recent and growing body of literature that proposes the framework of fractals to model optimization trajectories of neural networks, motivating generalization bounds and measures based on the fractal dimension of the trajectory. Notably, the persistent homology dimension has been proposed to correlate with the generalization gap. This paper performs an empirical evaluation of these persistent homology-based generalization measures, with an in-depth statistical analysis. Our study reveals confounding effects in the observed correlation between generalization and topological measures due to the variation of hyperparameters. We also observe that fractal dimension fails to predict generalization of models trained from poor initializations. We lastly reveal the intriguing manifestation of model-wise double descent in these topological generalization measures. Our work forms a basis for a deeper investigation of the causal relationships between fractal geometry, topological data analysis, and neural network optimization.
title On the Limitations of Fractal Dimension as a Measure of Generalization
topic Machine Learning
Artificial Intelligence
Dynamical Systems
url https://arxiv.org/abs/2406.02234