mlsanityData to 7 Oct 2026

Improved Masked Image Generation with Token-Critic

arXiv 2209.04439 ↗First posted 9 Sep 2022cs.CVcited by 63

Not on the map. Why?

Cited by 21 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Non-autoregressive generative transformers recently demonstrated impressive image generation performance, and orders of magnitude faster sampling than their autoregressive counterparts. However, optimal parallel sampling from the true joint distribution of visual tokens remains an open challenge. In this paper we introduce Token-Critic, an auxiliary model to guide the sampling of a non-autoregressive generative transformer. Given a masked-and-reconstructed real image, the Token-Critic model is trained to distinguish which visual tokens belong to the original image and which were sampled by the generative transformer. During non-autoregressive iterative sampling, Token-Critic is used to select which tokens to accept and which to reject and resample. Coupled with Token-Critic, a state-of-the-art generative transformer significantly improves its performance, and outperforms recent diffusion models and GANs in terms of the trade-off between generated image quality and diversity, in the challenging class-conditional ImageNet generation.

Key facts. Cited by 63 papers in the map (data through 7 Oct 2026). Builds on 36 of its 41 references that are in the map. First posted on arXiv 9 Sep 2022.

Builds on

36 of 41 references.

References · 36

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  4. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  5. 2020Language Models are Few-Shot Learnerscited by 32,199
  6. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  7. 2020Denoising Diffusion Implicit Modelscited by 9,655
  8. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  9. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  10. 2016Layer Normalizationcited by 5,579
  11. 2017Neural Discrete Representation Learningcited by 4,650
  12. 2021Zero-Shot Text-to-Image Generationcited by 4,276
24 more
  1. 2013Auto-Encoding Variational Bayescited by 3,980
  2. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
  3. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  4. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
  5. 2016Improved Techniques for Training GANscited by 2,938
  6. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  7. 2015Variational Inference with Normalizing Flowscited by 2,161
  8. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
  9. 2019Generating Diverse High-Fidelity Images with VQ-VAE-2cited by 1,376
  10. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,155
  11. 2021Variational Diffusion Modelscited by 1,145
  12. 2022MaskGIT: Masked Generative Image Transformercited by 1,082
  13. 2019Improved Precision and Recall Metric for Assessing Generative Modelscited by 925
  14. 2021Vector Quantized Diffusion Model for Text-to-Image Synthesiscited by 773
  15. 2021Vector-quantized Image Modeling with Improved VQGANcited by 656
  16. 2020NVAE: A Deep Hierarchical Variational Autoencodercited by 635
  17. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 566
  18. 2022StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasetscited by 489
  19. 2021ImageBART: Bidirectional Context with Multinomial Diffusion for Autoregressive Image Synthesiscited by 122
  20. 2020Fully Non-autoregressive Neural Machine Translation: Tricks of the Tradecited by 99
  21. 2021BLT: Bidirectional Layout Transformer for Controllable Layout Generationcited by 69
  22. 2021M6-UFC: Unifying Multi-Modal Controls for Conditional Image Synthesis via Non-Autoregressive Generative Transformerscited by 36
  23. 2019Discrete Object Generation with Reversible Inductive Constructioncited by 16
  24. 2020Incorporating a Local Translation Mechanism into Non-autoregressive Translationcited by 9

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 14; 2024: 20; 2025: 16; 2026: 10.

63 papers.

Sort:

By area:

  1. 2023Muse: Text-To-Image Generation via Masked Generative Transformerscited by 640
  2. 2023Language Model Beats Diffusion -- Tokenizer is Key to Visual Generationcited by 633
  3. 2022MAGVIT: Masked Generative Video Transformercited by 341
  4. 2024An Image is Worth 32 Tokens for Reconstruction and Generationcited by 295
  5. 2024NaturalSpeech 3: Zero-Shot Speech Synthesis with Factorized Codec and Diffusion Modelscited by 222
  6. 2024MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformercited by 159
  7. 2023A Reparameterized Discrete Diffusion Model for Text Generationcited by 152
  8. 2023Copilot4D: Learning Unsupervised World Models for Autonomous Driving via Discrete Diffusioncited by 105
  9. 2024MaskBit: Embedding-free Image Generation via Bit Tokenscited by 96
  10. 2022Visual Prompt Tuning for Generative Transfer Learningcited by 75
  11. 2023Zero-shot spatial layout conditioning for text-to-image diffusion modelscited by 64
  12. 2024Informed Correctors for Discrete Diffusion Modelscited by 57
38 more
  1. 2024Masked Audio Generation using a Single Non-Autoregressive Transformercited by 50
  2. 2023Vector Quantized Time Series Generation with a Bidirectional Prior Modelcited by 49
  3. 2025Fine-Tuning Masked Diffusion for Provable Self-Correctioncited by 39
  4. 2022Image Compression with Product Quantized Masked Image Modelingcited by 32
  5. 2025Halton Scheduler For Masked Generative Image Transformercited by 30
  6. 2024STAR: Scale-wise Text-conditioned AutoRegressive image generationcited by 28
  7. 2024Masked Image Modeling: A Surveycited by 28
  8. 2023MaskSketch: Unpaired Structure-guided Masked Image Generationcited by 27
  9. 2025Next Block Prediction: Video Generation via Semi-Autoregressive Modelingcited by 26
  10. 2023A Pytorch Reproduction of Masked Generative Image Transformercited by 25
  11. 2025AnyEnhance: A Unified Generative Model with Prompt-Guidance and Self-Critic for Voice Enhancementcited by 23
  12. 2024Revisiting Non-Autoregressive Transformers for Efficient Image Synthesiscited by 22
  13. 2024σ-GPTs: A New Approach to Autoregressive Modelscited by 20
  14. 2025Iterative Predictor-Critic Code Decoding for Real-World Image Dehazingcited by 16
  15. 2024AdaNAT: Exploring Adaptive Policy for Token-Based Image Generationcited by 13
  16. 2025TaDiCodec: Text-aware Diffusion Speech Tokenizer for Speech Language Modelingcited by 12
  17. 2023StraIT: Non-autoregressive Generation with Stratified Image Transformercited by 10
  18. 2025Resurrect Mask AutoRegressive Modeling for Efficient and Scalable Image Generationcited by 9
  19. 2023Learning Disentangled Prompts for Compositional Image Synthesiscited by 8
  20. 2024ECHOPulse: ECG controlled echocardio-grams video generationcited by 8
  21. 2024ENAT: Rethinking Spatial-temporal Interactions in Token-based Image Synthesiscited by 8
  22. 2024Layout-Corrector: Alleviating Layout Sticking Phenomenon in Discrete Diffusion Modelcited by 7
  23. 2024MaskMamba: A Hybrid Mamba-Transformer Model for Masked Image Generationcited by 7
  24. 2023Text-Conditioned Sampling Framework for Text-to-Image Generation with Masked Generative Modelscited by 6
  25. 2024Joint Semantic Knowledge Distillation and Masked Acoustic Modeling for Full-band Speech Restoration with Improved Intelligibilitycited by 5
  26. 2024Improving Vector-Quantized Image Modeling with Latent Consistency-Matching Diffusioncited by 5
  27. 2025Plug-and-Play Context Feature Reuse for Efficient Masked Generationcited by 5
  28. 2026Denoising, Fast and Slow: Difficulty-Aware Adaptive Sampling for Image Generationcited by 4
  29. 2023You Can Mask More For Extremely Low-Bitrate Image Compressioncited by 3
  30. 2023Masked Generative Modeling with Enhanced Sampling Schemecited by 3
  31. 2024Unlocking the Capabilities of Masked Generative Models for Image Synthesis via Self-Guidancecited by 3
  32. 2025Layton: Latent Consistency Tokenizer for 1024-pixel Image Reconstruction and Generation by 256 Tokenscited by 3
  33. 2025ReDDiT: Rehashing Noise for Discrete Visual Generationcited by 3
  34. 2025A Semantic Information-based Hierarchical Speech Enhancement Method Using Factorized Codec and Diffusion Modelcited by 2
  35. 2026SceneNAT: Masked Generative Modeling for Language-Guided Indoor Scene Synthesiscited by 2
  36. 2024Music Proofreading with RefinPaint: Where and How to Modify Compositions given Contextcited by 1
  37. 2024Dynamic Motion Synthesis: Masked Audio-Text Conditioned Spatio-Temporal Transformerscited by 1
  38. 2025Improved Masked Image Generation with Knowledge-Augmented Token Representationscited by 1

50 most cited of 63 All 63 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024High-Resolution Image Synthesis via Next-Token Prediction
  2. 2024Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation
  3. 2024Bag of Design Choices for Inference of High-Resolution Masked Generative Transformer
  4. 2021Unleashing Transformers: Parallel Token Prediction with Discrete Absorbing Diffusion for Fast High-Resolution Image Generation from Vector-Quantized Codes
  5. 2025Hierarchical Masked Autoregressive Models with Low-Resolution Token Pivots
  6. 2026Generative Refinement Networks for Visual Synthesis
  7. 2025Flow to the Mode: Mode-Seeking Diffusion Autoencoders for State-of-the-Art Image Tokenization
  8. 2024Efficient Generative Modeling with Residual Vector Quantization-Based Tokens