Survey of Diffusion Model Papers

  • Diffusion models: A comprehensive survey of methods and applications(202209 arXiv)[Paper] [Code]

    • 北大和斯坦福等高校联合写的一篇综述,主要从DM模型改进、DM应用以及DM与其他生成模型结合三个角度进行汇总分析
  • A Survey on Generative Diffusion Model(202209 arXiv)[Paper] [Code]

    • 这篇github上论文整理的挺好的

本文结合上述两篇综述对diffusion model相关论文进行整理

*表示一年以上google scholar引用量100+或者一年以内50+ (统计时间20221015)

次级标题进行补充说明 标题前的()是文章受关注后的别称

Diffusion Models Methods &Methods Improvement

  • Denoising Diffusion Probabilistic Models (2020 NurIPS) [Paper] [Code] *

  • Progressive distillation for fast sampling of diffusion models (2022 ICLR). [Paper] [Code] *

  • Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed (202101 arXiv) [Paper] [Code]

  • High-Resolution Image Synthesis with Latent Diffusion Models (2022 CVPR) [Paper] [Code] *

  • Diffusion Models Beat GANs on Image Synthesis (2021 NIPS) [Paper] [Code] *

  • Accelerating Diffusion Models via Early Stop of the Diffusion Process( 202205 arXiv) [Paper] [Code]

  • Truncated Diffusion Probabilistic Models (202202 arXiv) [Paper]

  • Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise(202208 arXiv) [Paper] [Code]

  • How Much is Enough? A Study on Diffusion Times in Score-based Generative Models (202206 arXiv) [Paper]

  • Improved Denoising Diffusion Probabilistic Models (2021 PMLR )[Paper] [Code] *

  • Noise estimation for generative diffusion models (202104 arXiv) [Paper]

  • Fast Sampling of Diffusion Models with Exponential Integrator (202204 arXiv) [Paper] [Code]

  • Variational diffusion models (2021 NIPS) [Paper] [Code] *

  • Structured denoising diffusion models in discrete state-spaces (2021 NIPS) [Paper] [Code] *

  • Denoising Diffusion Implicit Models ( ICLR 2020) [Paper] [Code] *

  • gDDIM: Generalized denoising diffusion implicit models (202206 arXiv) . [Paper] [Code]

  • Maximum Likelihood Training of Implicit Nonlinear Diffusion Models (202205 arXiv). [Paper]

  • Fast Sampling of Diffusion Models with Exponential Integrator (202204 arXiv).[Paper] [Code]

  • Pseudo numerical methods for diffusion models on manifolds (ICLR 2022) [Paper] [Code]

  • DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps (202206 arXiv)[Paper] [Code]

  • Gotta Go Fast When Generating Data with Score-Based Models (202105 arXiv)[Paper] [Code]

  • Learning Fast Samplers for Diffusion Models by Differentiating Through Sample Quality (ICLR 2022) [Paper]

  • Learning to efficiently sample from diffusion probabilistic models (202106 arXiv). [Paper]

Combine with Other Generative Models

  • Tackling the generative learning trilemma with denoising diffusion gans (202112 arXiv). [Paper] [Code]

  • Diffusion-GAN: Training GANs with Diffusion (202206 arXiv). [Paper] [Code]

  • Accelerating Diffusion Models via Early Stop of the Diffusion Process (202205 arXiv) [Paper] [Code]

  • Truncated diffusion probabilistic models (202202 arXiv) [Paper] [Code]

  • DiffuseVAE: Efficient, Controllable and High-Fidelity Generation from Low-Dimensional Latents (202201 arXiv) [Paper] [Code]

  • Diffusion normalizing flow**(NPIS 2021)** [Paper] [Code]

  • Maximum Likelihood Training of Implicit Nonlinear Diffusion Models (202205 arXiv). [Paper]

  • Score-based generative modeling in latent space (2021 NIPS). [Paper] [Code] *

  • Maximum likelihood training of score-based diffusion models (2021 NIPS) [Paper] [Code]

  • Maximum Likelihood Training of Parametrized Diffusion Model (2022 ICLR) [Paper]

  • Understanding Diffusion Models: A Unified Perspective (202208 arXiv) [Paper]

  • A variational perspective on diffusion-based generative models and score matching (2021 NIPS) [Paper]

  • Autoregressive Diffusion Models (2022 ICLR) . [Paper]

  • Autoregressive Denoising Diffusion Models for Multivariate Probabilistic Time Series Forecasting (2021 ICML). [Paper]

  • Learning Energy-Based Models by Diffusion Recovery Likelihood (2021 ICLR) [Paper] [Code]

  • Latent Diffusion Energy-Based Model for Interpretable Text Modeling (2022 ICML) [Paper] [Code]

Application of Diffusion Models

Computer Vision

Low-level Vision

  • Palette: Image-to-image diffusion models (2022 ACM SIGGRAPH) [ Paper] *
  • Denoising Diffusion Restoration Models (2022 ICLR) [Paper]
  • Srdiff: Single image super-resolution with diffusion probabilistic models (Neurocomputing 2022).
  • Image Super-Resolution via Iterative Refinement (2022 TPMI) [Paper] [Code]
  • Repaint: Inpainting using denoising diffusion probabilistic models (2022 CVPR) [Paper] [Code]

High-level Vision

  • Score-based generative modeling in latent space(2021 NIPS). [Paper] [Code]
  • Few-Shot Diffusion Models (202205 arXiv) [Paper]
  • GLIDE: Towards photorealistic image generation and editing with text-guided diffusion models(202112 arXiv) [Paper] [Code] *
  • (DALLE) Zero-Shot Text-to-Image Generation (2021 ICML) [Paper] [Code] *
  • (DALLE·2) Hierarchical text-conditional image generation with clip latents (202204 arXiv) [Paper] [Code] *
  • (Imagen) Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding(202205 arXiv)[Paper] [Code] *
  • DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation (202208 arXiv)[Paper] [Code]
  • ProDiff: Progressive Fast Diffusion Model For High-Quality Text-to-Speech (2022 ACM MM). [Paper] [Code]
  • Diffusion Causal Models for Counterfactual Estimation (2022 CLeaR) . [Paper] [Code]

3D Vision

  • Diffusion probabilistic models for 3d point cloud generation (2021 CVPR). [Paper] [Code] *
  • A conditional point diffusion-refinement paradigm for 3d point cloud completion (202112 arXiv). [Paper] [Code]
  • 3D Shape Generation and Completion Through Point-Voxel Diffusion (2021 ICCV) [Paper]
  • Score-based point cloud denoising (2021 ICCV) [Paper] [Code]

Video Modeling

  • Video Diffusion Models**( ICLR Workshop 2022)** [Paper] [Code]
  • Diffusion probabilistic modeling for video generation (202203 arXiv) [Paper] [Code] (code 待完善)
  • Flexible diffusion modeling of long videos (202205 arXiv). [Paper] [Code]
  • MCVD: Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation (202205 arXiv). [Paper] [Code]
  • Diffusion models for video prediction and infilling (202206 arXiv) [Paper]

Medical Application

  • Score-based diffusion models for accelerated MRI 2022 MedIA [Paper] [Code]
  • Solving Inverse Problems in Medical Imaging with Score-Based Generative Models (ICLR 2021). [Paper] [Code]
  • MR Image Denoising and Super-Resolution Using Regularized Reverse Diffusion (202206 arXiv) [Paper]
  • What is Healthy? Generative Counterfactual Diffusion for Lesion Localization (2022 MICCAI ) [Paper] [Code]

Sequential Modeling

Natural Language Processing

  • Diffusion-LM Improves Controllable Text Generation (202205 arXiv) [Paper] [Code]

  • Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning (202208 arXiv) [Paper]

  • Structured denoising diffusion models in discrete state-spacesp (2021 NIPS) [Paper]

Time Series

  • CSDI: Conditional score-based diffusion models for probabilistic time series imputation (2021 NIPS) [Paper] [Code]
  • Diffusion-based Time Series Imputation and Forecasting with Structured State Space Models. (202208 arXiv) [Paper] [Code]
  • Neural Markov Controlled SDE: Stochastic Optimization for Continuous-Time Data (2022 ICLR) [Paper]

Audio

Sound Generation

  • DiffWave: A Versatile Diffusion Model for Audio Synthesis (2020 ICLR) [Paper] [Code] *
  • WaveGrad: Estimating Gradients for Waveform Generation (2021 ICLR) [Paper] [Code] *
  • Grad-tts: A diffusion probabilistic model for text-to-speech (2021 ICML) [Paper] [Code]
  • Diff-TTS: A Denoising Diffusion Model for Text-to-Speech ( Interspeech 2021) [Paper]
  • Diffusion-Based Voice Conversion with Fast Maximum Likelihood Sampling Scheme (2021 ICLR) [Paper] [Code]
  • Diffsinger: Singing voice synthesis via shallow diffusion mechanism (AAAI 2022) [Paper] [Code]
  • Diffsound: Discrete Diffusion Model for Text-to-sound Generation (202207 arXiv) [Paper]

Text to Speech

  • EdiTTS: Score-based Editing for Controllable Text-to-Speech (202110 arXiv) [Paper] [Code]
  • Guided-TTS: A Diffusion Model for Text-to-Speech via Classifier Guidance (2022 ICML) [Paper]
  • Guided-TTS 2: A Diffusion Model for High-quality Adaptive Text-to-Speech with Untranscribed Data (202205 arXiv) [Paper]
  • Zero-Shot Voice Conditioning for Denoising Diffusion TTS Models (202206 arXiv) [Paper]
  • SpecGrad: Diffusion Probabilistic Model based Neural Vocoder with Adaptive Noise Spectral Shaping (202203 arXiv) [Paper]
  • BinauralGrad: A Two-Stage Conditional Diffusion Probabilistic Model for Binaural Audio Synthesis (202205 arXiv) [Paper]