mlsanityData to 7 Oct 2026

PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass Estimation

arXiv 2402.04355 ↗First posted 6 Feb 2024stat.ML cs.AI cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

We propose a likelihood-free method for comparing two distributions given samples from each, with the goal of assessing the quality of generative models. The proposed approach, PQMass, provides a statistically rigorous method for assessing the performance of a single generative model or the comparison of multiple competing models. PQMass divides the sample space into non-overlapping regions and applies chi-squared tests to the number of data samples that fall within each region, giving a p-value that measures the probability that the bin counts derived from two sets of samples are drawn from the same multinomial distribution. PQMass does not depend on assumptions regarding the density of the true distribution, nor does it rely on training or fitting any auxiliary models. We evaluate PQMass on data of various modalities and dimensions, demonstrating its effectiveness in assessing the quality, novelty, and diversity of generated samples. We further show that PQMass scales well to moderately high-dimensional data and thus obviates the need for feature extraction in practical applications.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 45 of its 89 references that are in the map. First posted on arXiv 6 Feb 2024.

Builds on

45 of 89 references.

References · 45

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  3. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  4. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  5. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  6. 2013Auto-Encoding Variational Bayescited by 3,980
  7. 2016Improved Techniques for Training GANscited by 2,938
  8. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,277
  9. 2015Variational Inference with Normalizing Flowscited by 2,161
  10. 2022Diffusion Posterior Sampling for General Noisy Inverse Problemscited by 1,375
  11. 2019Normalizing Flows for Probabilistic Modeling and Inferencecited by 1,315
  12. 2022Denoising Diffusion Restoration Modelscited by 951
33 more
  1. 2019Modeling Tabular data using Conditional GANcited by 695
  2. 2021Solving Inverse Problems in Medical Imaging with Score-Based Generative Modelscited by 514
  3. 2016f-GAN: Training Generative Neural Samplers using Variational Divergence Minimizationcited by 504
  4. 2019The frontier of simulation-based inferencecited by 490
  5. 2018Do Deep Generative Models Know What They Don't Know?cited by 443
  6. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 418
  7. 2018Assessing Generative Models via Precision and Recallcited by 401
  8. 2017Deep Models Under the GAN: Information Leakage from Collaborative Deep Learningcited by 326
  9. 2021Score-based diffusion models for accelerated MRIcited by 308
  10. 2015A note on the evaluation of generative modelscited by 291
  11. 2019Financial Time Series Forecasting with Deep Learning : A Systematic Literature Review: 2005-2019cited by 270
  12. 2018Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flowscited by 217
  13. 2019Automatic Posterior Transformation for Likelihood-Free Inferencecited by 216
  14. 2023Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion modelscited by 199
  15. 2021How Faithful is your Synthetic Data? Sample-level Metrics for Evaluating and Auditing Generative Modelscited by 193
  16. 2021Path Integral Sampler: a stochastic control approach for samplingcited by 175
  17. 2019Time-Series Anomaly Detection Service at Microsoftcited by 163
  18. 2023Score-Based Diffusion Models as Principled Priors for Inverse Imagingcited by 126
  19. 2022Flow Annealed Importance Sampling Bootstrapcited by 124
  20. 2018Improving GANs Using Optimal Transportcited by 122
  21. 2023Improved sampling via learned diffusionscited by 103
  22. 2023A theory of continuous generative flow networkscited by 98
  23. 2023Score-based Data Assimilationcited by 91
  24. 2023PFGM++: Unlocking the Potential of Physics-Inspired Generative Modelscited by 73
  25. 2023Sampling-Based Accuracy Testing of Posterior Estimators for General Inferencecited by 45
  26. 2020Perfect density models cannot guarantee anomaly detectioncited by 39
  27. 2023Bayesian Synthetic Likelihoodcited by 29
  28. 2022Probabilistic Mass Mapping with Neural Score Estimationcited by 17
  29. 2023Improving Gradient-guided Nested Sampling for Posterior Inferencecited by 14
  30. 2020Empirical Analysis of Overfitting and Mode Drop in GAN Trainingcited by 10
  31. 2023Diffusion Models for Probabilistic Deconvolution of Galaxy Imagescited by 4
  32. 2023Bayesian Imaging for Radio Interferometry with Score-Based Priorscited by 3
  33. 2023Echoes in the Noise: Posterior Samples of Faint Galaxy Surface Brightness Profiles with Score-Based Likelihoods and Priorscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 5; 2026: 3.

8 papers.

  1. 2025Random Initialization Can't Catch Up: The Advantage of Language Model Transfer for Time Series Forecastingcited by 2
  2. 2025A Data-Driven Prism: Multi-View Source Separation with Diffusion Model Priorscited by 2
  3. 2025Blind Strong Gravitational Lensing Inversion: Joint Inference of Source and Lens Mass with Score-Based Modelscited by 2
  4. 2025Detecting Localized Density Anomalies in Multivariate Data via Coin-Flip Statisticscited by 1
  5. 2025PRIVET: PRoximIty leakage detection Via Extreme value Theorycited by 1
  6. 2026MIRA: A Score for Conditional Distribution Accuracy and Model Comparisoncited by 1
  7. 2026An Introduction to Bayesian and Frequentist Simulation-Based Inference with Machine Learningcited by 1
  8. 2026PTED: A multi-dimensional two-sample test for scientific inference and generative machine learningcited by 0