(Deep) Generative Geodesics
Not on the map. Why?
Abstract (arXiv)
In this work, we propose to study the global geometrical properties of generative models. We introduce a new Riemannian metric to assess the similarity between any two data points. Importantly, our metric is agnostic to the parametrization of the generative model and requires only the evaluation of its data likelihood. Moreover, the metric leads to the conceptual definition of generative distances and generative geodesics, whose computation can be done efficiently in the data space. Their approximations are proven to converge to their true values under mild conditions. We showcase three proof-of-concept applications of this global metric, including clustering, data visualization, and data interpolation, thus providing new tools to support the geometrical understanding of generative models.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 19 of its 29 references that are in the map. First posted on arXiv 15 Jul 2024.
Builds on
19 of 29 references.
References · 19
- 2016Density estimation using Real NVPcited by 2,042
- 2023Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusioncited by 459
- 2017Latent Space Oddity: on the Curvature of Deep Generative Modelscited by 174
- 2023Understanding the Latent Space of Diffusion Models through the Lens of Riemannian Geometrycited by 125
- 2018Understanding and Improving Interpolation in Autoencoders via an Adversarial Regularizercited by 88
- 2017The Riemannian Geometry of Deep Generative Modelscited by 82
- 2017Metrics for Deep Generative Modelscited by 53
- 2021Low-Rank Subspaces in GANscited by 53
- 2020Geometrically Enriched Latent Spacescited by 45
- 2018A Spectral Regularizer for Unsupervised Disentanglementcited by 30
- 2021Do Not Escape From the Manifold: Discovering the Local Coordinates on the Latent Space of GANscited by 26
- 2020Learning Disentangled Representations with Latent Variation Predictabilitycited by 19
7 more
- 2023Latent Traversals in Generative Models as Potential Flowscited by 17
- 2023On Explicit Curvature Regularization in Deep Generative Modelscited by 13
- 2021A prior-based approximate latent Riemannian metriccited by 11
- 2021On Linear Interpolation in the Latent Space of Deep Generative Modelscited by 8
- 2017Approximating geodesics via random pointscited by 6
- 2019Feature-Based Interpolation and Geodesics in the Latent Spaces of Generative Modelscited by 5
- 2022Unveiling the Latent Space Geometry of Push-Forward Generative Modelscited by 3
Built on by
Citing papers per year: 2025: 1.
1 paper.