mlsanityData to 7 Oct 2026

STANLEY: Stochastic Gradient Anisotropic Langevin Dynamics for Learning Energy-Based Models

arXiv 2310.12667 ↗First posted 19 Oct 2023stat.ML cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We propose in this paper, STANLEY, a STochastic gradient ANisotropic LangEvin dYnamics, for sampling high dimensional data. With the growing efficacy and potential of Energy-Based modeling, also known as non-normalized probabilistic modeling, for modeling a generative process of different natures of high dimensional data observations, we present an end-to-end learning algorithm for Energy-Based models (EBM) with the purpose of improving the quality of the resulting sampled data points. While the unknown normalizing constant of EBMs makes the training procedure intractable, resorting to Markov Chain Monte Carlo (MCMC) is in general a viable option. Realizing what MCMC entails for the EBM training, we propose in this paper, a novel high dimensional sampling method, based on an anisotropic stepsize and a gradient-informed covariance matrix, embedded into a discretized Langevin diffusion. We motivate the necessity for an anisotropic update of the negative samples in the Markov Chain by the nonlinearity of the backbone of the EBM, here a Convolutional Neural Network. Our resulting method, namely STANLEY, is an optimization algorithm for training Energy-Based models via our newly introduced MCMC method. We provide a theoretical understanding of our sampling scheme by proving that the sampler leads to a geometrically uniformly ergodic Markov Chain. Several image generation experiments are provided in our paper to show the effectiveness of our method.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 39 of its 89 references that are in the map. First posted on arXiv 19 Oct 2023.

Builds on

39 of 89 references.

References · 39

  1. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  2. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,462
  3. 2020Energy-based Out-of-distribution Detectioncited by 1,296
  4. 2017Reinforcement Learning with Deep Energy-Based Policiescited by 659
  5. 2019Sliced Score Matching: A Scalable Approach to Density and Score Estimationcited by 381
  6. 2016A Theory of Generative ConvNetcited by 174
  7. 2019Learning Non-Convergent Non-Persistent Short-Run MCMC Toward Energy-Based Modelcited by 147
  8. 2020Improved Contrastive Divergence Training of Energy Based Modelscited by 127
  9. 2019On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Modelscited by 105
  10. 2020Learning Latent Space Energy-Based Prior Modelcited by 104
  11. 2020Residual Energy-Based Models for Text Generationcited by 95
  12. 2019Flow Contrastive Estimation of Energy-Based Modelscited by 75
27 more
  1. 2016Cooperative Training of Descriptor and Generator Networkscited by 60
  2. 2020Generative PointNet: Deep Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction and Classificationcited by 57
  3. 2021Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Predictioncited by 57
  4. 2018Learning Descriptor Networks for 3D Shape Synthesis and Analysiscited by 56
  5. 2022A Tale of Two Flows: Cooperative Learning of Langevin Flow and Normalizing Flow Toward Energy-Based Modelcited by 44
  6. 2020Learning Energy-Based Model with Variational Auto-Encoder as Amortized Samplercited by 43
  7. 2018Learning deep kernels for exponential family densitiescited by 36
  8. 2018Divergence Triangle for Joint Training of Generator Model, Energy-based Model, and Inference Modelcited by 34
  9. 2017Learning Energy-Based Models as Generative ConvNets via Multi-grid Modeling and Samplingcited by 32
  10. 2017The Tamed Unadjusted Langevin Algorithmcited by 31
  11. 2020Joint Training of Variational Auto-Encoder and Latent Energy-Based Modelcited by 30
  12. 2020Training Deep Energy-Based Models with f-Divergence Minimizationcited by 29
  13. 2020Learning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency Detectioncited by 29
  14. 2020Generative VoxelNet: Learning Energy-Based Models for 3D Shape Synthesis and Analysiscited by 26
  15. 2019Learning Energy-based Spatial-Temporal Generative ConvNets for Dynamic Patternscited by 22
  16. 2020Energy-based models for atomic-resolution protein conformationscited by 20
  17. 2017Wasserstein Introspective Neural Networkscited by 18
  18. 2018Learning Dynamic Generator Model by Alternating Back-Propagation Through Timecited by 18
  19. 2019Energy-Based Continuous Inverse Optimal Controlcited by 14
  20. 2021Learning Cycle-Consistent Cooperative Networks via Alternating MCMC Teaching for Unsupervised Cross-Domain Translationcited by 8
  21. 2021Energy-Based Generative Cooperative Saliency Predictioncited by 8
  22. 2013Variational MCMCcited by 7
  23. 2023A Tale of Two Latent Flows: Learning Latent Space Normalizing Flow with Short-run Langevin Flow for Approximate Inferencecited by 7
  24. 2022An Energy-Based Prior for Generative Saliencycited by 3
  25. 2023CoopInit: Initializing Generative Adversarial Networks via Cooperative Learningcited by 2
  26. 2023Likelihood-Based Generative Radiance Field with Latent Space Energy-Based Model for 3D-Aware Disentangled Image Representationcited by 2
  27. 2023Progressive Energy-Based Cooperative Learning for Multi-Domain Image-to-Image Translationcited by 1

Built on by

Not cited yet.