mlsanityData to 7 Oct 2026

Does Unsupervised Domain Adaptation Improve the Robustness of Amortized Bayesian Inference? A Systematic Evaluation

arXiv 2502.04949 ↗First posted 7 Feb 2025stat.ML cs.LG stat.MEcited by 10

Not on the map. Why?

Abstract (arXiv)

Neural networks are fragile when confronted with data that significantly deviates from their training distribution. This is true in particular for simulation-based inference methods, such as neural amortized Bayesian inference (ABI), where models trained on simulated data are deployed on noisy real-world observations. Recent robust approaches employ unsupervised domain adaptation (UDA) to match the embedding spaces of simulated and observed data. However, the lack of comprehensive evaluations across different domain mismatches raises concerns about the reliability in high-stakes applications. We address this gap by systematically testing UDA approaches across a wide range of misspecification scenarios in silico and practice. We demonstrate that aligning summary spaces between domains effectively mitigates the impact of unmodeled phenomena or noise. However, the same alignment mechanism can lead to failures under prior misspecifications - a critical finding with practical consequences. Our results underscore the need for careful consideration of misspecification types when using UDA to increase the robustness of ABI.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 59 of its 104 references that are in the map. First posted on arXiv 7 Feb 2025.

Builds on

59 of 104 references.

References · 59

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  2. 2022Flow Matching for Generative Modelingcited by 5,426
  3. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  4. 2015Domain-Adversarial Training of Neural Networkscited by 3,023
  5. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,455
  6. 2018Glow: Generative Flow with Invertible 1x1 Convolutionscited by 1,539
  7. 2017Deep Setscited by 1,394
  8. 2015Learning Transferable Features with Deep Adaptation Networkscited by 1,323
  9. 2019Normalizing Flows for Probabilistic Modeling and Inferencecited by 1,315
  10. 2021Domain Generalization: A Surveycited by 760
  11. 2014Deep Domain Confusion: Maximizing for Domain Invariancecited by 597
  12. 2019The frontier of simulation-based inferencecited by 490
38 more
  1. 2016Kernel Mean Embedding of Distributions: A Review and Beyondcited by 360
  2. 2018Analyzing Inverse Problems with Invertible Neural Networkscited by 239
  3. 2018Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flowscited by 217
  4. 2013A General Framework for Updating Belief Distributionscited by 172
  5. 2020BayesFlow: Learning complex stochastic models with invertible neural networkscited by 139
  6. 2023Modern Bayesian Experimental Designcited by 121
  7. 2019Unsupervised Domain Adaptation through Self-Supervisioncited by 116
  8. 2020A survey on domain adaptation theory: learning bounds and theoretical guaranteescited by 109
  9. 2022Deep Unsupervised Domain Adaptation: A Review of Recent Advances and Perspectivescited by 98
  10. 2023Flow Matching for Scalable Simulation-Based Inferencecited by 97
  11. 2018Unsupervised domain adaptation for medical imaging segmentation with self-ensemblingcited by 94
  12. 2019Support and Invertibility in Domain-Invariant Representationscited by 87
  13. 2020Neural Approximate Sufficient Statistics for Implicit Modelscited by 72
  14. 2024All-in-one simulation-based inferencecited by 71
  15. 2021Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networkscited by 48
  16. 2023Learning Robust Statistics for Simulation-based Inference under Model Misspecificationcited by 48
  17. 2020Bayesian Workflowcited by 47
  18. 2022Robust Neural Posterior Estimation and Statistical Model Criticismcited by 45
  19. 2022Investigating the Impact of Model Misspecification in Neural Simulation-based Inferencecited by 44
  20. 2022Variational methods for simulation-based inferencecited by 43
  21. 2022GATSBI: Generative Adversarial Training for Simulation-Based Inferencecited by 36
  22. 2024Neural Methods for Amortized Inferencecited by 36
  23. 2023Free Lunch for Domain Adversarial Training: Environment Label Smoothingcited by 33
  24. 2023BayesFlow: Amortized Bayesian Workflows With Neural Networkscited by 32
  25. 2016Inference Compilation and Universal Probabilistic Programmingcited by 31
  26. 2018A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networkscited by 30
  27. 2024Addressing Misspecification in Simulation-based Inference through Data-driven Calibrationcited by 26
  28. 2022Reliable amortized variational inference with physics-based latent distribution correctioncited by 25
  29. 2021Conditional Invertible Neural Networks for Diverse Image-to-Image Translationcited by 24
  30. 2021Unsupervised Domain Adaptation: A Reality Checkcited by 24
  31. 2016Deep Amortized Inference for Probabilistic Programscited by 22
  32. 2020Amortized Bayesian model comparison with evidential deep learningcited by 22
  33. 2023Sensitivity-Aware Amortized Bayesian Inferencecited by 20
  34. 2023Adversarial robustness of amortized Bayesian inferencecited by 17
  35. 2023Generalized Bayesian Inference for Scientific Simulators via Amortized Cost Estimationcited by 17
  36. 2023DUSA: Decoupled Unsupervised Sim2Real Adaptation for Vehicle-to-Everything Collaborative Perceptioncited by 17
  37. 2022Some models are useful, but how do we know which ones? Towards a unified Bayesian model taxonomycited by 16
  38. 2024The Statistical Accuracy of Neural Posterior and Likelihood Estimationcited by 15

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 4; 2026: 6.

10 papers.

  1. 2025Multifidelity Simulation-based Inference for Computationally Expensive Simulatorscited by 9
  2. 2026BayesFlow 2: Multi-Backend Amortized Bayesian Inference in Pythoncited by 7
  3. 2025Inductive Domain Transfer In Misspecified Simulation-Based Inferencecited by 4
  4. 2026Minimum Distance Summaries for Robust Neural Posterior Estimationcited by 3
  5. 2025Uncertainty-Aware Surrogate-based Amortized Bayesian Inference for Computationally Expensive Modelscited by 1
  6. 2025Misspecification-robust amortised simulation-based inference using variational methodscited by 0
  7. 2026Unsupervised Continual Learning for Amortized Bayesian Inferencecited by 0
  8. 2026Overcoming Selection Bias in Statistical Studies With Amortized Bayesian Inferencecited by 0
  9. 2026Information-Preserving Domain Transfer with Unlabeled Data in Misspecified Simulation-Based Inferencecited by 0
  10. 2026Mitigating Representation Gaps in Amortized Bayesian Inference with Auxiliary Supervisioncited by 0