Predicting the generalization gap in neural networks using topological data analysis
This addresses the problem of understanding generalization in neural networks for researchers and practitioners, but it is incremental as it builds on existing methods with a novel application.
The paper tackled predicting the generalization gap in neural networks by using topological data analysis on neuron activation correlations, achieving competitive prediction accuracy on CIFAR10 and SVHN datasets without needing a test set.
Understanding how neural networks generalize on unseen data is crucial for designing more robust and reliable models. In this paper, we study the generalization gap of neural networks using methods from topological data analysis. For this purpose, we compute homological persistence diagrams of weighted graphs constructed from neuron activation correlations after a training phase, aiming to capture patterns that are linked to the generalization capacity of the network. We compare the usefulness of different numerical summaries from persistence diagrams and show that a combination of some of them can accurately predict and partially explain the generalization gap without the need of a test set. Evaluation on two computer vision recognition tasks (CIFAR10 and SVHN) shows competitive generalization gap prediction when compared against state-of-the-art methods.