Geometry Score: A Method For Comparing Generative Adversarial Networks
This addresses the problem of evaluating GAN performance for researchers, offering a more robust metric beyond visual inspection, though it is incremental as it builds on existing evaluation challenges.
The paper tackles the challenge of evaluating generative adversarial networks (GANs) by proposing a method that compares geometrical properties of real and generated data manifolds to assess sample quality and detect mode collapse. The method works on arbitrary datasets and provides both qualitative and quantitative insights, as demonstrated on various real-life models and datasets.
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.