mlsanityData to 14 Oct 2026

Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

arXiv 2105.08164 ↗First posted 17 May 2021cs.LG cs.SD eess.AScited by 19

Not on the map. Why?

Abstract (arXiv)

This paper introduces an alternative approach to sampling from autoregressive models. Autoregressive models are typically sampled sequentially, according to the transition dynamics defined by the model. Instead, we propose a sampling procedure that initializes a sequence with white noise and follows a Markov chain defined by Langevin dynamics on the global log-likelihood of the sequence. This approach parallelizes the sampling process and generalizes to conditional sampling. Using an autoregressive model as a Bayesian prior, we can steer the output of a generative model using a conditional likelihood or constraints. We apply these techniques to autoregressive models in the visual and audio domains, with competitive results for audio source separation, super-resolution, and inpainting.

Key facts. Cited by 19 papers in the map (data through 14 Oct 2026). Builds on 39 of its 62 references that are in the map. First posted on arXiv 17 May 2021.

Builds on

39 of 62 references.

References · 39

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  3. 2017Neural Discrete Representation Learningcited by 4,712
  4. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
  5. 2019The Curious Case of Neural Text Degenerationcited by 2,916
  6. 2016WaveNet: A Generative Model for Raw Audiocited by 1,722
  7. 2019Generating Long Sequences with Sparse Transformerscited by 1,546
  8. 2016Context Encoders: Feature Learning by Inpaintingcited by 1,428
  9. 2019Generating Diverse High-Fidelity Images with VQ-VAE-2cited by 1,387
  10. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesiscited by 1,354
  11. 2020Improved Techniques for Training Score-Based Generative Modelscited by 1,043
  12. 2016Pixel Recurrent Neural Networkscited by 816
27 more
  1. 2018Residual Dense Network for Image Super-Resolutioncited by 799
  2. 2018Image Transformercited by 772
  3. 2020WaveGrad: Estimating Gradients for Waveform Generationcited by 622
  4. 2020Jukebox: A Generative Model for Musiccited by 542
  5. 2018SDR - half-baked or well done?cited by 423
  6. 2017PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modificationscited by 417
  7. 2018Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separationcited by 406
  8. 2019MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesiscited by 354
  9. 2017Compressed Sensing using Generative Modelscited by 339
  10. 2018WaveGlow: A Flow-based Generative Network for Speech Synthesiscited by 309
  11. 2018Adversarial Audio Synthesiscited by 225
  12. 2017Parallel WaveNet: Fast High-Fidelity Speech Synthesiscited by 215
  13. 2016SampleRNN: An Unconditional End-to-End Neural Audio Generation Modelcited by 110
  14. 2019Music Source Separation in the Waveform Domaincited by 99
  15. 2016DeepBach: a Steerable Model for Bach Chorales Generationcited by 91
  16. 2018ClariNet: Parallel Wave Generation in End-to-End Text-to-Speechcited by 70
  17. 2017One Network to Solve Them All --- Solving Linear Inverse Problems using Deep Projection Modelscited by 67
  18. 2013RNADE: The real-valued neural autoregressive density-estimatorcited by 66
  19. 2018FloWaveNet : A Generative Flow for Raw Audiocited by 54
  20. 2019WaveFlow: A Compact Flow-based Model for Raw Audiocited by 51
  21. 2017Pixel Recursive Super Resolutioncited by 44
  22. 2020Source Separation with Deep Generative Priorscited by 36
  23. 2015Generative Image Modeling Using Spatial LSTMscited by 35
  24. 2021Anytime Sampling for Autoregressive Models via Ordered Autoencodingcited by 11
  25. 2020Predictive Sampling with Forecasting Autoregressive Modelscited by 10
  26. 2018Neural separation of observed and unobserved distributionscited by 6
  27. 2020Problems using deep generative models for probabilistic audio source separationcited by 2

Built on by

Cited by year

Citing papers per year: 2021: 2; 2022: 7; 2023: 5; 2024: 4; 2025: 1.

19 papers.

Sort:

By area:

  1. 2022Continuous diffusion for categorical datacited by 214
  2. 2022It's Raw! Audio Generation with State-Space Modelscited by 171
  3. 2022COLD Decoding: Energy-based Constrained Text Generation with Langevin Dynamicscited by 161
  4. 2022ShapeFormer: Transformer-based Shape Completion via Sparse Representationcited by 106
  5. 2022Gradient-Based Constrained Sampling from Language Modelscited by 64
  6. 2023Multi-Source Diffusion Models for Simultaneous Music Generation and Separationcited by 63
  7. 2021Multimodal Image Synthesis and Editing: The Generative AI Eracited by 52
  8. 2022Diffusion-based Generative Speech Source Separationcited by 41
  9. 2023AI-Generated Images as Data Source: The Dawn of Synthetic Eracited by 13
  10. 2024Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Modelscited by 13
  11. 2023Latent Autoregressive Source Separationcited by 12
  12. 2021Unsupervised Source Separation via Bayesian Inference in the Latent Domaincited by 2
7 more
  1. 2025Efficient and Fast Generative-Based Singing Voice Separation using a Latent Diffusion Modelcited by 2
  2. 2022Autoregressive Generative Modeling with Noise Conditional Maximum Likelihood Estimationcited by 1
  3. 2023Distribution Preserving Source Separation With Time Frequency Predictive Modelscited by 1
  4. 2024Audio Decoding by Inverse Problem Solvingcited by 1
  5. 2024Improving Music Source Separation with Diffusion and Consistency Refinementcited by 1
  6. 2023Interpretable ODE-style Generative Diffusion Model via Force Field Constructioncited by 0
  7. 2024Source Separation of Multi-source Raw Music using a Residual Quantized Variational Autoencodercited by 0