Data to 14 Oct 2026

Latent Diffusion Energy-Based Model for Interpretable Text Modeling

arXiv 2206.05895 ↗First posted 13 Jun 2022cs.LG cs.CLcited by 67

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

Latent space Energy-Based Models (EBMs), also known as energy-based priors, have drawn growing interests in generative modeling. Fueled by its flexibility in the formulation and strong modeling power of the latent space, recent works built upon it have made interesting attempts aiming at the interpretability of text modeling. However, latent space EBMs also inherit some flaws from EBMs in data space; the degenerate MCMC sampling quality in practice can lead to poor generation quality and instability in training, especially on data with complex latent structures. Inspired by the recent efforts that leverage diffusion recovery likelihood learning as a cure for the sampling issue, we introduce a novel symbiosis between the diffusion models and latent space EBMs in a variational learning framework, coined as the latent diffusion energy-based model. We develop a geometric clustering-based regularization jointly with the information bottleneck to further improve the quality of the learned latent space. Experiments on several challenging tasks demonstrate the superior performance of our model on interpretable text modeling over strong counterparts.

Key facts. Cited by 67 papers in the map (data through 14 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 58 of its 80 references that are in the map. First posted on arXiv 13 Jun 2022.

Builds on

58 of 80 references.

References · 58

2014Adam: A Method for Stochastic Optimizationcited by 37,240
2020Denoising Diffusion Probabilistic Modelscited by 21,076
2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
2013Auto-Encoding Variational Bayescited by 4,053
2017Billion-scale similarity search with GPUscited by 2,497
2015Character-level Convolutional Networks for Text Classificationcited by 2,330
2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
2018Spectral Normalization for Generative Adversarial Networkscited by 1,636
2020Improved Techniques for Training Score-Based Generative Modelscited by 1,043

38 more

2017DailyDialog: A Manually Labelled Multi-turn Dialogue Datasetcited by 652
2020NVAE: A Deep Hierarchical Variational Autoencodercited by 637
2021Score-based Generative Modeling in Latent Spacecited by 636
2015Generating Sentences from a Continuous Spacecited by 612
2014Semi-Supervised Learning with Deep Generative Modelscited by 496
2015Importance Weighted Autoencoderscited by 456
2019Your Classifier is Secretly an Energy Based Model and You Should Treat it Like Onecited by 405
2017Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoderscited by 222
2015Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Modelscited by 207
2017Regularizing and Optimizing LSTM Language Modelscited by 205
2018Delete, Retrieve, Generate: A Simple Approach to Sentiment and Style Transfercited by 191
2019Cyclical Annealing Schedule: A Simple Approach to Mitigating KL Vanishingcited by 191
2016A Theory of Generative ConvNetcited by 174
2016A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialoguescited by 174
2019Learning Non-Convergent Non-Persistent Short-Run MCMC Toward Energy-Based Modelcited by 148
2017Key-Value Retrieval Networks for Task-Oriented Dialoguecited by 138
2020Improved Contrastive Divergence Training of Energy Based Modelscited by 130
2019Lagging Inference Networks and Posterior Collapse in Variational Autoencoderscited by 124
2020Learning Energy-Based Models by Diffusion Recovery Likelihoodcited by 123
2013Generalized Denoising Auto-Encoders as Generative Modelscited by 113
2015Neural Variational Inference for Text Processingcited by 111
2019On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Modelscited by 105
2020Learning Latent Space Energy-Based Prior Modelcited by 104
2021D2C: Diffusion-Denoising Models for Few-shot Conditional Generationcited by 96
2021Controllable and Compositional Generation with Latent-Space Energy-Based Modelscited by 64
2021Trajectory Prediction with Latent Belief Energy-Based Modelcited by 60
2017A Deep Generative Framework for Paraphrase Generationcited by 54
2017Neural Models for Documents with Metadatacited by 46
2019Topic-Guided Variational Autoencoders for Text Generationcited by 44
2019Variational Pretraining for Semi-supervised Text Classificationcited by 41
2016Variational Neural Machine Translationcited by 35
2018Unsupervised Discrete Sentence Representation Learning for Interpretable Neural Dialog Generationcited by 34
2018Divergence Triangle for Joint Training of Generator Model, Energy-based Model, and Inference Modelcited by 34
2019A Surprisingly Effective Fix for Deep Latent Variable Modeling of Textcited by 32
2020Joint Training of Variational Auto-Encoder and Latent Energy-Based Modelcited by 30
2019Implicit Deep Latent Variable Models for Text Generationcited by 26
2021Unsupervised Foreground Extraction via Deep Region Competitioncited by 26
2021Diffusion Priors In Variational Autoencoderscited by 25

50 most cited of 58 All 58 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 7; 2023: 23; 2024: 25; 2025: 9; 2026: 3.

67 papers.

Sort:

By area:

2022Diffusion Models: A Comprehensive Survey of Methods and Applicationscited by 1,114
2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Modelscited by 450
2023Diffusion-based Generation, Optimization, and Planning in 3D Scenescited by 239
2024PhyScene: Physically Interactable 3D Scene Synthesis for Embodied AIcited by 125
2025A Survey on Diffusion Language Modelscited by 78
2022Diffusion models for missing value imputation in tabular datacited by 77
2023Diffusion Action Segmentationcited by 74
2024An Overview of Diffusion Models: Applications, Guided Generation, Statistical Rates and Optimizationcited by 73
2023Diffusion Models for Time Series Applications: A Surveycited by 72
2022SeqDiffuSeq: Text Diffusion with Encoder-Decoder Transformerscited by 64
2023Enhancing Deep Reinforcement Learning: A Tutorial on Generative Diffusion Models in Network Optimizationcited by 56
2024Flow Priors for Linear Inverse Problems via Iterative Corrupted Trajectory Matchingcited by 53

38 more

2023DINOISER: Diffused Conditional Sequence Learning by Manipulating Noisescited by 50
2022DiffusionInst: Diffusion Model for Instance Segmentationcited by 49
2023RenAIssance: A Survey into AI Text-to-Image Generation in the Era of Large Modelcited by 47
2025Diffusion Models for Robotic Manipulation: A Surveycited by 41
2023An Ensemble Score Filter for Tracking High-Dimensional Nonlinear Dynamical Systemscited by 38
2024Attacks and Defenses for Generative Diffusion Models: A Comprehensive Surveycited by 32
2022Composable Text Controls in Latent Space with ODEscited by 28
2024The Rise of Diffusion Models in Time-Series Forecastingcited by 25
2025Latent Thought Models with Variational Bayes Inference-Time Computationcited by 25
2023Distribution Shift Inversion for Out-of-Distribution Predictioncited by 23
2023Quantum-Noise-Driven Generative Diffusion Modelscited by 20
2024A Training-Free Conditional Diffusion Model for Learning Stochastic Dynamical Systemscited by 19
2022Adaptive Multi-stage Density Ratio Estimation for Learning Latent Space Energy-based Modelcited by 18
2024Object-Conditioned Energy-Based Attention Map Alignment in Text-to-Image Diffusion Modelscited by 18
2023Diffusion Models in NLP: A Surveycited by 16
2023Diffusion Models in Bioinformatics: A New Wave of Deep Learning Revolution in Actioncited by 15
2025Wasserstein Convergence of Score-based Generative Models under Semiconvexity and Discontinuous Gradientscited by 15
2023Diffusion-SS3D: Diffusion Model for Semi-supervised 3D Object Detectioncited by 14
2023Learning Energy-Based Prior Model with Diffusion-Amortized MCMCcited by 13
2023Diffusion-Model-Assisted Supervised Learning of Generative Models for Density Estimationcited by 13
2023MIRACLE: Towards Personalized Dialogue Generation with Latent-Space Multiple Personal Attribute Controlcited by 11
2024TimeLDM: Latent Diffusion Model for Unconditional Time Series Generationcited by 10
2023DiffAnt: Diffusion Models for Action Anticipationcited by 9
2024DiffLM: Controllable Synthetic Data Generation via Diffusion Language Modelscited by 9
2023A Comprehensive Survey on Generative Diffusion Models for Structured Datacited by 8
2023Molecule Design by Latent Space Energy-Based Modeling and Gradual Distribution Shiftingcited by 8
2024"Noisier" Noise Contrastive Eestimation is (Almost) Maximum Likelihoodcited by 8
2024Active Diffusion Subsamplingcited by 8
2025An efficient probabilistic hardware architecture for diffusion-like modelscited by 8
2024Watermarking Generative Tabular Datacited by 7
2026Streaming-dLLM: Accelerating Diffusion LLMs via Suffix Pruning and Dynamic Decodingcited by 7
2024Learning Latent Space Hierarchical EBM Diffusion Modelscited by 6
2024SpecSTG: A Fast Spectral Diffusion Framework for Probabilistic Spatio-Temporal Traffic Forecastingcited by 5
2024Hitchhiker's guide on the relation of Energy-Based Models with other generative models, sampling and statistical physics: a comprehensive reviewcited by 5
2025SketchDNN: Joint Continuous-Discrete Diffusion for CAD Sketch Generationcited by 5
2023DiffuseGAE: Controllable and High-fidelity Image Manipulation from Disentangled Representationcited by 4
2024Language Rectified Flow: Advancing Diffusion Language Generation with Probabilistic Flowscited by 4
2024Latent Space Energy-based Neural ODEscited by 4

50 most cited of 67 All 67 (JSON)

Similar, not cited

Close in wording, no citation either way