mlsanityData to 14 Oct 2026

Diagnosing and Enhancing VAE Models

arXiv 1903.05789 ↗First posted 14 Mar 2019cs.LG cs.CV stat.MLcited by 193

Not on the map. Why?

Cited by 31 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Although variational autoencoders (VAEs) represent a widely influential deep generative model, many aspects of the underlying energy function remain poorly understood. In particular, it is commonly believed that Gaussian encoder/decoder assumptions reduce the effectiveness of VAEs in generating realistic samples. In this regard, we rigorously analyze the VAE objective, differentiating situations where this belief is and is not actually true. We then leverage the corresponding insights to develop a simple VAE enhancement that requires no additional hyperparameters or sensitive tuning. Quantitatively, this proposal produces crisp samples and stable FID scores that are actually competitive with a variety of GAN models, all while retaining desirable attributes of the original VAE architecture. A shorter version of this work will appear in the ICLR 2019 conference proceedings (Dai and Wipf, 2019). The code for our model is available at https://github.com/daib13/ TwoStageVAE.

Key facts. Cited by 193 papers in the map (data through 14 Oct 2026). First posted on arXiv 14 Mar 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 43; 2022: 49; 2023: 39; 2024: 22; 2025: 25; 2026: 15.

193 papers.

Sort:

By area:

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2021Score-based Generative Modeling in Latent Spacecited by 636
  3. 2021Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Modelscited by 252
  4. 2023Geometric Latent Diffusion Models for 3D Molecule Generationcited by 186
  5. 2021ImageBART: Bidirectional Context with Multinomial Diffusion for Autoregressive Image Synthesiscited by 122
  6. 2021SynFace: Face Recognition with Synthetic Datacited by 106
  7. 2022DiffuseVAE: Efficient, Controllable and High-Fidelity Generation from Low-Dimensional Latentscited by 104
  8. 2021Geometry-Free View Synthesis: Transformers and no 3D Priorscited by 100
  9. 2021D2C: Diffusion-Denoising Models for Few-shot Conditional Generationcited by 96
  10. 2022Verifying the Union of Manifolds Hypothesis for Image Datacited by 64
  11. 2023Adaptive Latent Diffusion Model for 3D Medical Image to Image Translation: Multi-modal Magnetic Resonance Imaging Studycited by 64
  12. 2021Soft then Hard: Rethinking the Quantization in Neural Image Compressioncited by 51
38 more
  1. 2021Rectangular Flows for Manifold Learningcited by 50
  2. 2023ControlMat: A Controlled Generative Approach to Material Capturecited by 50
  3. 2021Generating Synthetic Mixed-type Longitudinal Electronic Health Records for Artificial Intelligent Applicationscited by 47
  4. 2024Deep Generative Models through the Lens of the Manifold Hypothesis: A Survey and New Connectionscited by 47
  5. 2022A Tale of Two Flows: Cooperative Learning of Langevin Flow and Normalizing Flow Toward Energy-Based Modelcited by 44
  6. 2024A Geometric Framework for Understanding Memorization in Generative Modelscited by 40
  7. 2022Recent Advances for Quantum Neural Networks in Generative Learningcited by 39
  8. 2023Generative Sliced MMD Flows with Riesz Kernelscited by 37
  9. 2021Tractable Density Estimation on Learned Manifolds with Conformal Embedding Flowscited by 31
  10. 2022Direct Molecular Conformation Generationcited by 30
  11. 2022AnoViT: Unsupervised Anomaly Detection and Localization with Vision Transformer-based Encoder-Decodercited by 29
  12. 2022Diagnosing and Fixing Manifold Overfitting in Deep Generative Modelscited by 28
  13. 2022Embrace the Gap: VAEs Perform Independent Mechanism Analysiscited by 28
  14. 2022Synthetic Data in Human Analysis: A Surveycited by 28
  15. 2021Data Augmentation in High Dimensional Low Sample Size Setting Using a Geometry-Based Variational Autoencodercited by 27
  16. 2022DiVAE: Photorealistic Images Synthesis with Denoising Diffusion Decodercited by 27
  17. 2022Particle algorithms for maximum likelihood training of latent variable modelscited by 26
  18. 2022Pythae: Unifying Generative Autoencoders in Python -- A Benchmarking Use Casecited by 26
  19. 2021Solving Inverse Problems by Joint Posterior Maximization with Autoencoding Priorcited by 24
  20. 2022Exact Solutions of a Deep Linear Networkcited by 24
  21. 2022A Geometric Perspective on Variational Autoencoderscited by 24
  22. 2021Generalized Normalizing Flows via Markov Chainscited by 23
  23. 2021A survey on Variational Autoencoders from a GreenAI perspectivecited by 22
  24. 2021Regularising Inverse Problems with Generative Machine Learning Modelscited by 22
  25. 20223DFaceShop: Explicitly Controllable 3D-Aware Portrait Generationcited by 22
  26. 2023Learning Manifold Dimensions with Conditional Variational Autoencoderscited by 20
  27. 2021Sentence Bottleneck Autoencoders from Transformer Language Modelscited by 19
  28. 2022Continuous Mixtures of Tractable Probabilistic Modelscited by 19
  29. 2023Complexity Matters: Rethinking the Latent Space for Generative Modelingcited by 19
  30. 2022Adaptive Multi-stage Density Ratio Estimation for Learning Latent Space Energy-based Modelcited by 18
  31. 2022NVDiff: Graph Generation through the Diffusion of Node Vectorscited by 18
  32. 2022Gaussian-Bernoulli RBMs Without Tearscited by 17
  33. 2023Manifold Learning by Mixture Models of VAEs for Inverse Problemscited by 17
  34. 2021Preventing Oversmoothing in VAE via Generalized Variance Parameterizationcited by 15
  35. 2023Canonical normalizing flows for manifold learningcited by 15
  36. 2021Cauchy-Schwarz Regularized Autoencodercited by 13
  37. 2023Learning Energy-Based Prior Model with Diffusion-Amortized MCMCcited by 13
  38. 2024Geometric Representation Condition Improves Equivariant Molecule Generationcited by 13

50 most cited of 193 All 193 (JSON)