Geometry Score: A Method For Comparing Generative Adversarial Networks

Abstract

One of the biggest challenges in the research of generative adversarial networks (GANs) is assessing the quality of generated samples and detecting various levels of mode collapse. In this work, we construct a novel measure of performance of a GAN by comparing geometrical properties of the underlying data manifold and the generated one, which provides both qualitative and quantitative means for evaluation. Our algorithm can be applied to datasets of an arbitrary nature and is not limited to visual data. We test the obtained metric on various real-life models and datasets and demonstrate that our method provides new insights into properties of GANs.

ICML 2018

References in corpus (3)

• CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks
• A Note on the Inception Score
• The Riemannian Geometry of Deep Generative Models

Cited by in corpus (19)

• Classification Accuracy Score for Conditional Generative Models
• Neural Persistence: A Complexity Measure for Deep Neural Networks Using Algebraic Topology
• CorrGAN: Sampling Realistic Financial Correlation Matrices Using Generative Adversarial Networks
• Uncovering the Topology of Time-Varying fMRI Data using Cubical Persistence
• The Shape of Data: Intrinsic Distance for Data Distributions
• Evaluating the Disentanglement of Deep Generative Models through Manifold Topology
• Constrained Generative Adversarial Network Ensembles for Sharable Synthetic Data Generation
• MR-GAN: Manifold Regularized Generative Adversarial Networks
• Generating artificial digital image correlation data using physics-guided adversarial networks
• Universality Theorems for Generative Models
• Manifold Topology Divergence: a Framework for Comparing Data Manifolds
• Multivariate-Information Adversarial Ensemble for Scalable Joint Distribution Matching
• Improving the Evaluation of Generative Models with Fuzzy Logic
• Diffusion models for Handwriting Generation
• Double Descent and Other Interpolation Phenomena in GANs
• Rethinking conditional GAN training: An approach using geometrically structured latent manifolds
• A Novel Measure to Evaluate Generative Adversarial Networks Based on Direct Analysis of Generated Images
• Towards Interaction Detection Using Topological Analysis on Neural Networks
• Toward a Generalization Metric for Deep Generative Models