Topological Generalization Bounds for Discrete-Time Stochastic Optimization Algorithms

TL;DR

This paper introduces new topology-based complexity measures for discrete-time training trajectories of deep neural networks, providing reliable bounds on generalization error that outperform existing methods across various architectures and datasets.

Contribution

It develops computationally efficient, topology-based complexity measures tailored for discrete-time training data, offering provable bounds on generalization error without restrictive assumptions.

Findings

Complexity measures strongly correlate with generalization error.

Outperforms existing topological bounds on multiple datasets and architectures.

Applicable to diverse domains, tasks, and neural network architectures.

Abstract

We present a novel set of rigorous and computationally efficient topology-based complexity notions that exhibit a strong correlation with the generalization gap in modern deep neural networks (DNNs). DNNs show remarkable generalization properties, yet the source of these capabilities remains elusive, defying the established statistical learning theory. Recent studies have revealed that properties of training trajectories can be indicative of generalization. Building on this insight, state-of-the-art methods have leveraged the topology of these trajectories, particularly their fractal dimension, to quantify generalization. Most existing works compute this quantity by assuming continuous- or infinite-time training dynamics, complicating the development of practical estimators capable of accurately predicting generalization without access to test data. In this paper, we respect the discrete-time nature of training trajectories and investigate the underlying topological quantities that can be amenable to topological data analysis tools. This leads to a new family of reliable topological complexity measures that provably bound the generalization error, eliminating the need for restrictive geometric assumptions. These measures are computationally friendly, enabling us to propose simple yet effective algorithms for computing generalization indices. Moreover, our flexible framework can be extended to different domains, tasks, and architectures. Our experimental results demonstrate that our new complexity measures correlate highly with generalization error in industry-standards architectures such as transformers and deep graph networks. Our approach consistently outperforms existing topological bounds across a wide range of datasets, models, and optimizers, highlighting the practical relevance and effectiveness of our complexity measures.