mlsanityData to 7 Oct 2026

Iterative Amortized Inference

arXiv 1807.09356 ↗First posted 24 Jul 2018cs.LG stat.MLcited by 93

Not on the map. Why?

Abstract (arXiv)

Inference models are a key component in scaling variational inference to deep latent variable models, most notably as encoder networks in variational auto-encoders (VAEs). By replacing conventional optimization-based inference with a learned model, inference is amortized over data examples and therefore more computationally efficient. However, standard inference models are restricted to direct mappings from data to approximate posterior estimates. The failure of these models to reach fully optimized approximate posterior estimates results in an amortization gap. We aim toward closing this gap by proposing iterative inference models, which learn to perform inference optimization through repeatedly encoding gradients. Our approach generalizes standard inference models in VAEs and provides insight into several empirical findings, including top-down inference techniques. We demonstrate the inference optimization capabilities of iterative inference models and show that they outperform standard inference models on several benchmark data sets of images and text.

Key facts. Cited by 93 papers in the map (data through 7 Oct 2026). First posted on arXiv 24 Jul 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 17; 2022: 20; 2023: 13; 2024: 18; 2025: 8; 2026: 17.

93 papers.

Sort:

By area:

  1. 2021COIN: COmpression with Implicit Neural representationscited by 232
  2. 2022NeRF: Neural Radiance Field in 3D Vision: A Comprehensive Review (Updated Post-Gaussian Splatting)cited by 156
  3. 2021NeRF-VAE: A Geometry Aware 3D Scene Generative Modelcited by 111
  4. 2021Recent advances and clinical applications of deep learning in medical image analysiscited by 104
  5. 2022Tutorial on amortized optimizationcited by 83
  6. 2022Object Representations as Fixed Points: Training Iterative Refinement Algorithms with Implicit Differentiationcited by 46
  7. 2021Learning Object-Centric Representations of Multi-Object Scenes from Multiple Viewscited by 37
  8. 2021Efficient Iterative Amortized Inference for Learning Symmetric and Disentangled Multi-Object Representationscited by 36
  9. 2023Amortized Variational Inference: When and Why?cited by 28
  10. 2022Compositional Scene Representation Learning via Reconstruction: A Surveycited by 27
  11. 2021Unsupervised Foreground Extraction via Deep Region Competitioncited by 26
  12. 2022Hybrid Predictive Coding: Inferring, Fast and Slowcited by 25
38 more
  1. 2021Preconditioned training of normalizing flows for variational inference in inverse problemscited by 19
  2. 2022Flexible Neural Image Compression via Code Editingcited by 19
  3. 2022Amortized Projection Optimization for Sliced Wasserstein Generative Modelscited by 17
  4. 2023Grounded Object Centric Learningcited by 17
  5. 2024Compute Optimal Inference and Provable Amortisation Gap in Sparse Autoencoderscited by 17
  6. 2022A Theoretical Framework for Inference and Learning in Predictive Coding Networkscited by 16
  7. 2022Amortized Variational Inference: A Systematic Reviewcited by 16
  8. 2022Conditional Injective Flows for Bayesian Imagingcited by 13
  9. 2021Backprop-Free Reinforcement Learning with Active Neural Generative Codingcited by 12
  10. 2023Importance Weighted Expectation-Maximization for Protein Sequence Designcited by 12
  11. 2021Unsupervised Learning of Compositional Scene Representations from Multiple Unspecified Viewpointscited by 11
  12. 2024Memory-based Cross-modal Semantic Alignment Network for Radiology Report Generationcited by 11
  13. 2024ASPIRE: Iterative Amortized Posterior Inference for Bayesian Inverse Problemscited by 11
  14. 2024Poisson Variational Autoencodercited by 11
  15. 2021Latent Space Refinement for Deep Generative Modelscited by 10
  16. 2021Stochastic Iterative Graph Matchingcited by 10
  17. 2021Object-Centric Representation Learning with Generative Spatial-Temporal Factorizationcited by 10
  18. 2023Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedbackcited by 10
  19. 2023WISE: full-Waveform variational Inference via Subsurface Extensionscited by 10
  20. 2021Regularization-Agnostic Compressed Sensing MRI Reconstruction with Hypernetworkscited by 9
  21. 2022Bit Allocation using Optimizationcited by 9
  22. 2023Fully Bayesian Autoencoders with Latent Sparse Gaussian Processescited by 9
  23. 2024Identifiable Object-Centric Representation Learning via Probabilistic Slot Attentioncited by 9
  24. 2022Object-based active inferencecited by 7
  25. 2022Variational Laplace Autoencoderscited by 7
  26. 2021Reducing the Amortization Gap in Variational Autoencoders: A Bayesian Random Function Approachcited by 6
  27. 2021Knowledge-Guided Object Discovery with Acquired Deep Impressionscited by 6
  28. 2022Optimization for Amortized Inverse Problemscited by 6
  29. 2023Active Sensing with Predictive Coding and Uncertainty Minimizationcited by 5
  30. 2025Bidirectional predictive codingcited by 5
  31. 2021Unsupervised Representation Learning via Neural Activation Codingcited by 4
  32. 2022Thalamus: a brain-inspired algorithm for biologically-plausible continual learning and disentangled representationscited by 4
  33. 2022On the Relationship Between Variational Inference and Auto-Associative Memorycited by 4
  34. 2023Perceptual Group Tokenizer: Building Perception with Iterative Groupingcited by 4
  35. 2024Unsupervised Object-Centric Learning from Multiple Unspecified Viewpointscited by 4
  36. 2024How does the primate brain combine generative and discriminative computations in vision?cited by 4
  37. 2024Neural NeRF Compressioncited by 4
  38. 2024CryoSPIN: Improving Ab-Initio Cryo-EM Reconstruction with Semi-Amortized Pose Inferencecited by 4

50 most cited of 93 All 93 (JSON)