We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Latent Space Refinement for Deep Generative Models

Abstract: Deep generative models are becoming widely used across science and industry for a variety of purposes. A common challenge is achieving a precise implicit or explicit representation of the data probability density. Recent proposals have suggested using classifier weights to refine the learned density of deep generative models. We extend this idea to all types of generative models and show how latent space refinement via iterated generative modeling can circumvent topological obstructions and improve precision. This methodology also applies to cases were the target model is non-differentiable and has many internal latent dimensions which must be marginalized over before refinement. We demonstrate our Latent Space Refinement (LaSeR) protocol on a variety of examples, focusing on the combinations of Normalizing Flows and Generative Adversarial Networks.
Comments: 15 pages, 5 figures, 3 tables
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph); Data Analysis, Statistics and Probability (physics.data-an)
Journal reference: NeurIPS DGMs and Applications Workshop 2021
Report number: CP3-21-61
Cite as: arXiv:2106.00792 [stat.ML]
  (or arXiv:2106.00792v2 [stat.ML] for this version)

Submission history

From: Ramon Winterhalder [view email]
[v1] Tue, 1 Jun 2021 21:01:39 GMT (1478kb,D)
[v2] Wed, 3 Nov 2021 19:21:11 GMT (1486kb,D)

Link back to: arXiv, form interface, contact.