mlsanityData to 7 Oct 2026

Learning Robust Statistics for Simulation-based Inference under Model Misspecification

arXiv 2305.15871 ↗First posted 25 May 2023stat.ML cs.LG stat.COcited by 48

Not on the map. Why?

Abstract (arXiv)

Simulation-based inference (SBI) methods such as approximate Bayesian computation (ABC), synthetic likelihood, and neural posterior estimation (NPE) rely on simulating statistics to infer parameters of intractable likelihood models. However, such methods are known to yield untrustworthy and misleading inference outcomes under model misspecification, thus hindering their widespread applicability. In this work, we propose the first general approach to handle model misspecification that works across different classes of SBI methods. Leveraging the fact that the choice of statistics determines the degree of misspecification in SBI, we introduce a regularized loss function that penalises those statistics that increase the mismatch between the data and the model. Taking NPE and ABC as use cases, we demonstrate the superior performance of our method on high-dimensional time-series models that are artificially misspecified. We also apply our method to real data from the field of radio propagation where the model is known to be misspecified. We show empirically that the method yields robust inference in misspecified scenarios, whilst still being accurate when the model is well-specified.

Key facts. Cited by 48 papers in the map (data through 7 Oct 2026). Builds on 53 of its 87 references that are in the map. First posted on arXiv 25 May 2023.

Builds on

53 of 87 references.

References · 53

  1. 2017Deep Setscited by 1,394
  2. 2017Masked Autoregressive Flow for Density Estimationcited by 680
  3. 2019The frontier of simulation-based inferencecited by 490
  4. 2016Kernel Mean Embedding of Distributions: A Review and Beyondcited by 360
  5. 2018Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flowscited by 217
  6. 2019Automatic Posterior Transformation for Likelihood-Free Inferencecited by 216
  7. 2013A General Framework for Updating Belief Distributionscited by 172
  8. 2017Flexible statistical inference for mechanistic models of neural dynamicscited by 162
  9. 2020BayesFlow: Learning complex stochastic models with invertible neural networkscited by 139
  10. 2020On Contrastive Learning for Likelihood-free Inferencecited by 96
  11. 2016Likelihood-free inference by ratio estimationcited by 80
  12. 2020Neural Approximate Sufficient Statistics for Implicit Modelscited by 72
38 more
  1. 2018Likelihood-free inference with emulator networkscited by 61
  2. 2019Generalized Variational Inference: Three arguments for deriving new Posteriorscited by 57
  3. 2022Sequential Neural Score Estimation: Likelihood-Free Inference with Conditional Score Based Diffusion Modelscited by 50
  4. 2021Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networkscited by 48
  5. 2022Robust Neural Posterior Estimation and Statistical Model Criticismcited by 45
  6. 2022Investigating the Impact of Model Misspecification in Neural Simulation-based Inferencecited by 44
  7. 2019Statistical Inference for Generative Models with Maximum Mean Discrepancycited by 43
  8. 2022Variational methods for simulation-based inferencecited by 43
  9. 2022Robust Bayesian Inference for Simulator-based Models via the MMD Posterior Bootstrapcited by 41
  10. 2015Learning Summary Statistic for Approximate Bayesian Computation via Deep Neural Networkcited by 39
  11. 2019Conditional Density Estimation with Neural Networks: Best Practices and Benchmarkscited by 37
  12. 2022GATSBI: Generative Adversarial Training for Simulation-Based Inferencecited by 36
  13. 2019MMD-Bayes: Robust Bayesian Estimation via Maximum Mean Discrepancycited by 35
  14. 2018A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networkscited by 30
  15. 2021Sequential Neural Posterior and Likelihood Approximationcited by 30
  16. 2023Bayesian Synthetic Likelihoodcited by 29
  17. 2022Towards Reliable Simulation-Based Inference with Balanced Neural Ratio Estimationcited by 27
  18. 2020Complete parameter inference for GW150914 using deep learningcited by 25
  19. 2019Finite sample properties of parametric MMD estimation: robustness to misspecification and dependencecited by 24
  20. 2020Generalized Posteriors in Approximate Bayesian Computationcited by 19
  21. 2021Group equivariant neural posterior estimationcited by 16
  22. 2019Robust Approximate Bayesian Inference with Synthetic Likelihoodcited by 14
  23. 2022Likelihood-Free Inference with Generative Neural Networks via Scoring Rule Minimizationcited by 14
  24. 2022Black-box Bayesian inference for economic agent-based modelscited by 13
  25. 2023Optimally-Weighted Estimators of the Maximum Mean Discrepancy for Likelihood-Free Inferencecited by 13
  26. 2019Partially Exchangeable Networks and Architectures for Learning Summary Statistics in Approximate Bayesian Computationcited by 12
  27. 2021Discrepancy-based Inference for Intractable Generative Models using Quasi-Monte Carlocited by 12
  28. 2020A General Method for Calibrating Stochastic Radio Channel Models with Kernelscited by 9
  29. 2022Learning Summary Statistics for Bayesian Inference with Autoencoderscited by 9
  30. 2022Approximate Bayesian Computation with Domain Expert in the Loopcited by 9
  31. 2022Amortised Likelihood-free Inference for Expensive Time-series Simulators with Signatured Ratio Estimationcited by 8
  32. 2023Misspecification-robust Sequential Neural Likelihood for Simulation-based Inferencecited by 8
  33. 2018Dynamic Likelihood-free Inference via Ratio Estimation (DIRE)cited by 7
  34. 2021ABC of the Futurecited by 7
  35. 2020Bayesian Neural Networks With Maximum Mean Discrepancy Regularizationcited by 6
  36. 2021Bayesian data selectioncited by 6
  37. 2020Robust Approximate Bayesian Computation: An Adjustment Approachcited by 5
  38. 2021Synthetic likelihood in misspecified modelscited by 5

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 4; 2024: 12; 2025: 20; 2026: 12.

48 papers.

Sort:

By area:

  1. 2025Simulation-Based Inference: A Practical Guidecited by 30
  2. 2024Addressing Misspecification in Simulation-based Inference through Data-driven Calibrationcited by 26
  3. 2024Amortized Probabilistic Conditioning for Optimization, Simulation and Inferencecited by 22
  4. 2023Sensitivity-Aware Amortized Bayesian Inferencecited by 20
  5. 2024Amortized Bayesian Experimental Design for Decision-Makingcited by 20
  6. 2024Amortized Bayesian Multilevel Modelscited by 12
  7. 2025Robust Amortized Bayesian Inference with Self-Consistency Losses on Unlabeled Datacited by 11
  8. 2025Does Unsupervised Domain Adaptation Improve the Robustness of Amortized Bayesian Inference? A Systematic Evaluationcited by 10
  9. 2025Multifidelity Simulation-based Inference for Computationally Expensive Simulatorscited by 9
  10. 2025Simulation-based Bayesian inference under model misspecificationcited by 9
  11. 2025A Dictionary of Closed-Form Kernel Mean Embeddingscited by 9
  12. 2024Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networks: An Extended Investigationcited by 8
36 more
  1. 2025ALINE: Joint Amortization for Bayesian Inference and Active Data Acquisitioncited by 8
  2. 2025Solving Empirical Bayes via Transformerscited by 7
  3. 2026BayesFlow 2: Multi-Backend Amortized Bayesian Inference in Pythoncited by 7
  4. 2024Preconditioned Neural Posterior Estimation for Likelihood-free Inferencecited by 6
  5. 2024Cost-aware simulation-based inferencecited by 6
  6. 2024A Comprehensive Guide to Simulation-based Inference in Computational Biologycited by 5
  7. 2025Robust Simulation-Based Inference under Missing Data via Neural Processescited by 5
  8. 2023Deep Calibration of Market Simulations using Neural Density Estimators and Embedding Networkscited by 4
  9. 2025Inductive Domain Transfer In Misspecified Simulation-Based Inferencecited by 4
  10. 2023Fuse It or Lose It: Deep Fusion for Multimodal Simulation-Based Inferencecited by 3
  11. 2024A Quadrature Approach for General-Purpose Batch Bayesian Optimization via Probabilistic Liftingcited by 3
  12. 2024Amortized Bayesian Workflowcited by 3
  13. 2025Posterior SBC: Simulation-Based Calibration Checking Conditional on Datacited by 3
  14. 2025PriorGuide: Test-Time Prior Adaptation for Simulation-Based Inferencecited by 3
  15. 2026Minimum Distance Summaries for Robust Neural Posterior Estimationcited by 3
  16. 2025Cryo-em images are intrinsically low dimensionalcited by 2
  17. 2026Constrained Bayesian Experimental Design via Online Planningcited by 2
  18. 2023Simulation-based Inference for Exoplanet Atmospheric Retrieval: Insights from winning the Ariel Data Challenge 2023 using Normalizing Flowscited by 1
  19. 2024R'enyi Neural Processescited by 1
  20. 2024Sequential Kernelized Stein Discrepancycited by 1
  21. 2025Causal Posterior Estimationcited by 1
  22. 2025Representative, Informative, and De-Amplifying: Requirements for Robust Bayesian Active Learning under Model Misspecificationcited by 1
  23. 2025Simulation-based population inference of LISA's Galactic binaries: Bypassing the global fitcited by 1
  24. 2025Flow Matching Calibration for Simulation-Based Inference under Model Misspecificationcited by 1
  25. 2025Inferring Optical Tissue Properties from Photoplethysmography using Hybrid Amortized Inferencecited by 1
  26. 2025Deep Neural Network Driven Simulation Based Inference Method for Pole Position Estimation under Model Misspecificationcited by 0
  27. 2025Misspecification-robust amortised simulation-based inference using variational methodscited by 0
  28. 2026Questioning the Coverage-Length Metric in Conformal Prediction: When Shorter Intervals Are Not Bettercited by 0
  29. 2026Simulation-Based Inference via Regression Projection and Batched Discrepanciescited by 0
  30. 2026Preconditioned Robust Neural Posterior Estimation for Misspecified Simulatorscited by 0
  31. 2026Unsupervised Continual Learning for Amortized Bayesian Inferencecited by 0
  32. 2026Information-Preserving Domain Transfer with Unlabeled Data in Misspecified Simulation-Based Inferencecited by 0
  33. 2026Amortized Neural Clustering of Time Series based on Statistical Featurescited by 0
  34. 2026Benchmark of Likelihood-Free Inference Methods based on Neural and Optimal Transport Approachescited by 0
  35. 2026MA-SBI: Misspecification-Aware Simulation-Based Inference via Side-Channel Guidancecited by 0
  36. 2026Mitigating Representation Gaps in Amortized Bayesian Inference with Auxiliary Supervisioncited by 0