Pith. sign in

REVIEW 4 major objections 4 minor 46 references

Divide-and-Conquer: Towards Generalizable Amortized Bayesian Inference for the Drift Diffusion Model

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A single pairwise neural posterior estimator, applied to disjoint pairs of conditions and combined by consensus MCMC, can approximate the full drift-diffusion posterior for any experimental design.

desk verdict Useful divide-and-conquer ABI for DDM, but the 'exact' claim is overstated: the real-data boundary-separation gap shows the consensus step is approximate in practice. read the letter →

arxiv 2608.03566 v1 pith:FZU3HQAZ submitted 2026-08-04 stat.ML cs.LGstat.APstat.ME

Yufei Wu , Shanqing Gao , Andreas Voss , Francis Tuerlinckx This is my paper
classification stat.MLcs.LGstat.APstat.ME MSC 62F1562-0865C05
keywords driftdiffusionmodelamortizedBayesianinferenceneuralposteriorestimationconsensusMCMCconditionalflowmatchingdivide-and-conquerdesigngeneralizationcomputation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Amortized Bayesian inference (ABI) for the drift diffusion model (DDM) is nearly instantaneous, but a trained neural network is locked to the experimental design it saw during training. This paper proposes a divide-and-conquer fix: split any dataset into disjoint pairs of conditions, run one pairwise-trained network on every pair, and combine the pairwise posteriors through consensus MCMC with an importance-sampling correction. Because DDM trials are independent, the product of shard posteriors is mathematically equivalent to the full posterior, so the decomposition is not an approximation in principle. In two simulation settings, the calibrated consensus posterior matched full MCMC in both posterior means and uncertainties, with inference time in milliseconds rather than minutes. The real-data test showed the boundary of the method: when conditions shrink to about 10 trials, consensus estimates of boundary separation depart visibly from full MCMC, a failure the paper traces to non-Gaussian shard posteriors.

What carries the argument

The central mechanism is the exact factorization of the DDM likelihood over disjoint shards: p(x|θ) = ∏_q p(S_q|θ), which holds because trials are independent given the parameters. The paper combines this with a pairwise neural posterior estimator and consensus MCMC, where each shard's posterior draws are fused by a diagonal precision-weighted average (Eq. 2), and importance weights proportional to p(θ)^{1−Q} undo the prior overcounting introduced by analyzing Q shards separately. The Gaussian-product assumption in the combination step is what makes the whole pipeline tractable.

What would settle it

Simulate a dataset whose conditions are tiny (about 10 trials per shard), run the pairwise divide-and-conquer pipeline, and compare the calibrated consensus posterior to the full MCMC posterior; if the boundary-separation posterior mean and credible interval deviate beyond Monte Carlo error, the Gaussian consensus assumption is the failing link.

Watch

Extended reading notes

Core claim

The paper establishes that the DDM's independence across trials lets any experimental design be partitioned into pairwise shards that all share one common model structure: two conditions, two drift rates, and shared remaining parameters. A single neural posterior estimator trained only on two-condition data can therefore be applied to shards from any design, regardless of the number of conditions. The full posterior is recovered by multiplying the shard posteriors, correcting the Q-fold reuse of the prior with importance weights, and combining draws through a precision-weighted consensus average. The calibrated consensus posterior approximates the full posterior closely in simulations, demon

Load-bearing premise

The combination step assumes each shard posterior is close to Gaussian, so a diagonal precision-weighted average of posterior draws recovers the full posterior after importance weighting; with very small shards this premise fails, as the paper's 10-trial mini-block analysis shows.

Editorial extensions

If this is right

  • A single pairwise-trained network can serve any number of conditions without retraining, removing the main generalization bottleneck of ABI for DDM studies.
  • Inference time stays constant in dataset size and several orders of magnitude below MCMC; the paper reports roughly 730 ms for 100 datasets across six shards versus tens of minutes for MCMC.
  • The sharding strategy lowers the dimensionality of the inference problem, avoiding the biased posteriors the authors encountered when training high-dimensional full-model networks.
  • For MCMC, the same pairwise decomposition is not a computational win because each shard needs its own chain; the gains are specific to the amortized setting.
  • Accuracy and uncertainty of the consensus posterior approach those of full MCMC as trials per condition grow, with small residual over-dispersion in boundary separation parameters.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper, the same sharding trick should transfer to any generative model whose likelihood factorizes over independent units, such as hierarchical or multi-condition cognitive models beyond the DDM, provided a pairwise estimator can be trained for the smallest exchangeable unit.
  • The small-shard failure seen in the real-data analysis suggests a non-Gaussian combination step—for example, fitting a normalizing flow or a copula to the shard draws—could extend the framework to designs with very few trials per condition.
  • Because the pairwise network is trained on a broad range of trial counts, it could support adaptive experimental designs where condition sizes change mid-experiment, without retraining.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes a divide-and-conquer framework for amortized Bayesian inference (ABI) of the drift diffusion model (DDM). The full dataset is partitioned into pairwise condition shards, a single neural posterior estimator (NPE) trained on two-condition data is applied to each shard, and the shard posteriors are combined via consensus MCMC with an importance-sampling correction for prior over-counting. The authors claim that, because the DDM likelihood factorizes exactly over disjoint shards, the procedure is mathematically equivalent to full-data inference. They evaluate the method in two simulation studies (four-condition and 3×2 designs, n=100/200/500, 100 replicates each) with both ABI and MCMC, reporting accuracy (MAE, RMSE) and posterior SDs comparable to full-model inference, and orders-of-magnitude speedups for ABI. They also apply the method to the Ratcliff & Rouder (1998) dataset with 33×2 conditions, where they observe larger discrepancies from full MCMC, which they attribute to small (10-trial) mini-block shards producing non-Gaussian posteriors.

Significance. If the central claim holds, the paper offers a practical solution to a well-known limitation of amortized inference: design-specific NPEs do not generalize across experimental designs. The pairwise decomposition is an appealing idea that could extend ABI to arbitrary numbers of conditions without retraining, and the computational advantage over MCMC is substantial. The paper is also commendable for benchmarking against full-model MCMC, for including two distinct designs and multiple sample sizes, and for explicitly discussing limitations in the real-data example. However, the paper's strongest theoretical claim—that the implemented pipeline is 'mathematically equivalent to analyzing the full dataset at once'—is not supported for the actual consensus step. The discrepancy in the real-data analysis (Table 7 vs. Table A3) shows that the approximation is consequential in small-shard regimes, which is precisely the setting the method is advertised to handle. The contribution is valuable as an approximate and scalable inference strategy, but the exactness framing needs to be corrected and the approximation error characterized more carefully.

major comments (4)
  1. [§4, Eq. (2)] The statement in §4 that the decomposition is 'mathematically equivalent to analyzing the full dataset at once' is not supported by the implemented pipeline. The exact factorization applies to the product of exact shard posteriors p(θ|S_q), but Eq. (2) replaces each shard posterior with a Gaussian and then uses a diagonal precision-weighted average, explicitly forgoing the full covariance matrix. The consensus step is therefore approximate, not exact. The paper's own real-data analysis illustrates the consequence: for participant 1, a_accuracy is 1.653 (ABI) and 1.620 (MCMC) under consensus versus 2.092 for full MCMC (Table 7 vs. Table A3), and the authors attribute this to non-Gaussian 10-trial shard posteriors. This is the regime where Eq. (2)'s Gaussian/diagonal approximation is unjustified. Please revise the exactness claim to state precisely what is exact (the likelihood factorizati
  2. [§6, mini-block size] The real-data analysis uses 10-trial mini-blocks, chosen as the minimum trial count in rr98, but no sensitivity analysis is reported for this free parameter. The number of shards is very large (1056 pairs for participant 1), and the importance-sampling correction of Appendix B uses weights w(θ) ∝ p(θ)^{1-Q}. With Q on the order of hundreds, these weights can be extremely variable, and finite posterior draws (1000 per pair) may not reliably approximate the reweighted target. The paper does not report any diagnostic for the importance weights (e.g., effective sample size) or the stability of the calibrated consensus posterior. Please add such diagnostics and, at minimum, a small robustness check varying the mini-block size (e.g., 10 vs. 20 vs. 50 trials) to show how the gap relative to full MCMC depends on Q and shard size.
  3. [§5, Estimation; §6] The claim that 'a single pairwise NPE trained on pairwise data can be applied to any shard regardless of the original design' is subject to a range restriction that is acknowledged implicitly but not as a limitation. In the simulations, the pairwise NPE is trained on shard sizes U(30,300) trials per condition, yet the real data contains conditions with only 10 trials, so the authors retrain a separate NPE with training sizes 10–250. This means the design-universality claim only holds for shard sizes within the training range; the same network cannot be applied to arbitrarily small shards. Please state this explicitly and clarify that the real-data analysis uses a new pairwise NPE, not the one from the simulation studies.
  4. [Table 3; Table 7] The abstract's claim that the approach 'achieves accuracy and uncertainty comparable to MCMC' is supported by the simulation studies but is not supported by the real-data example, where the consensus estimates of a_accuracy differ from full MCMC by about 20% (1.653 vs. 2.092 for participant 1) and credible intervals are visibly wider and shifted at the tails of the brightness-strength range (Figure 14 vs. Figure A14). The paper's own text acknowledges this discrepancy. The conclusion should be moderated to distinguish the well-behaved large-shard regime from the small-shard regime, where the consensus approximation introduces nontrivial bias and overdispersion. As it stands, the abstract overstates the general empirical performance.
minor comments (4)
  1. [Section 2] Typographical issues: 'T er' appears in several places where 'Ter' is meant; '0 anda' should be '0 and a'; the diffusion constant notation σ is used but the parameter is not listed in the prior table. Please proofread.
  2. [Table 2] The caption says 'The bottom row reports the consensus estimate obtained from the relevant shards,' but the bottom row contains parameter name labels, not estimates. Clarify that the bottom row is the consensus parameter set, not numerical estimates.
  3. [Appendix B] The derivation of the importance weights is correct but terse. It would help to state explicitly that the draws from p(θ|x) in the first line are the consensus posterior draws, not the calibrated draws, and to define C and C* consistently. Also, 'θ⋆ l' in the text appears misindexed.
  4. [Section 5, Computational time] The sentence 'Obtaining 3,000 posterior draws from the full-model NPE required approximately 195 milliseconds for 100 simulated datasets' is ambiguous: 195 ms per dataset or total? From the context it seems per dataset, but please make it explicit, because the subsequent pairwise time of 730 ms is also per 100 datasets.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the divide-and-conquer decomposition rests on an exact likelihood factorization, and the acknowledged Gaussian/diagonal consensus approximation is a correctness limitation, not a self-referential prediction.

full rationale

The derivation chain is not circular. The core decomposition in Section 4 uses the DDM's conditional independence to write p(x|θ)=∏_q p(S_q|θ), an exact standard factorization, and the shard posteriors are defined as p(θ|S_q)∝p(S_q|θ)p(θ). The paper then derives that the full posterior is proportional to [∏_q p(θ|S_q)]·p(θ)^{1−Q} (Appendix B). Nothing in this chain fits parameters to the benchmark full posterior or renames the benchmark as a prediction. Consensus MCMC (Eq. 2) and importance sampling are imported from external prior work (Scott et al. 2022; Tokdar & Kass 2010), not from the authors, and the NPE is validated by parameter recovery and simulation-based calibration against simulation truth, not against the target conclusion. The only author self-citations (Voss & Voss 2007; Ratcliff & Tuerlinckx 2002) are routine DDM software/estimation citations and are not load-bearing. There is a genuine correctness limitation: the paper's statement that consensus MCMC with importance sampling is 'mathematically equivalent to analyzing the full dataset at once' is overstated because Eq. 2 replaces shard posteriors by a diagonal-Gaussian consensus, an approximation the paper itself acknowledges, and the real-data example shows a substantial gap (a_accuracy 1.653/1.620 vs 2.092, Table 7 vs Table A3) attributed to non-Gaussian 10-trial shard posteriors. However, overclaiming the accuracy of an approximate combination is a validity/calibration concern, not a circularity concern: the full posterior is computed independently by MCMC on the full data, and the approximation could have failed (and partially does), so the evaluation has independent content.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

Everything rests on the i.i.d. DDM assumption, the Gaussian/diagonal consensus approximation, the restricted pairwise parameterization, and extrapolation of the NPE to small shards. There are no newly invented entities. The only hand-chosen numeric input is the 10-trial mini-block size in the real-data example.

free parameters (1)
  • mini_block_size = 10 trials per mini-block
    Used in the rr98 real-data analysis to handle conditions with very few trials; chosen to match the minimum trial count, not derived from theory, and it directly affects shard size and the validity of the Gaussian consensus assumption.
assumptions (5)
  • domain assumption DDM observations are independent and identically distributed across trials and conditions given the parameters.
    Used in Section 4 for the exact factorization p(x|theta)=prod_q p(S_q|theta); the Discussion acknowledges this fails for sequential dependencies.
  • standard math Shard posteriors are approximately Gaussian (Bernstein-von Mises) at the sample sizes used.
    Invoked in Section 4 to justify product-of-Gaussians combination; the paper cites Le Cam 2012. Whether it holds for 10-trial mini-blocks is questionable.
  • ad hoc to paper The consensus step may use a diagonal covariance matrix W_q in Eq. 2 without losing accuracy.
    Section 4 says the optimal scenario uses the full covariance matrix but the paper approximates with a diagonal matrix, following Scott et al. 2022; no diagnostics for the lost correlations are reported.
  • domain assumption Each shard can be parameterized with two varying drift rates and shared a, z, Ter.
    Table 2 defines all shards this way; this restricts the 'any experimental design' claim to designs where non-drift parameters are shared within pairs or within instruction groups.
  • ad hoc to paper The pairwise NPE generalizes to shard sizes at or below its training range (U(30,300) trials per condition).
    The real-data example uses 10-trial mini-blocks, far below the training range; the paper assumes extrapolation works, with limited validation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Divide-and-Conquer: Towards Generalizable Amortized Bayesian Inference for the Drift Diffusion Model." pith.science (2026). https://pith.science/paper/FZU3HQAZ

@misc{pith2026260803566,
  author       = {Pith},
  title        = {Pith review of: Divide-and-Conquer: Towards Generalizable Amortized Bayesian Inference for the Drift Diffusion Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FZU3HQAZ}},
  note         = {Machine review of arXiv:2608.03566}
}
read the original abstract

The drift diffusion model (DDM) is a cornerstone of cognitive decision-making research. Although numerous estimation methods exist, researchers continue to seek inference approaches that are both fast and flexible across diverse study designs. Amortized Bayesian inference (ABI) can provide nearly instantaneous inference for complex stochastic models like the DDM, but neural networks trained for one study design cannot generalize to others. In this paper, we propose a divide-and-conquer framework that address this limitation. The core idea is that the DDM's independence assumption allows the full dataset to be decomposed into pairwise shards, each sharing a common structure that a single neural network can learn. Inference is performed on each shard separately and the resulting posteriors are combined via consensus MCMC to approximate the full posterior. Using simulated datasets, we evaluate the accuracy and uncertainty of this method. Our results show that the proposed divide-and-conquer approach achieves accuracy and uncertainty comparable to MCMC while reducing computational cost by several orders of magnitude. This work not only advances DDM estimation but also demonstrates a general strategy for improving the scalability and generalizability of ABI methods across diverse applications.

Figures

Figures reproduced from arXiv: 2608.03566 by the authors.

Figure 1
Figure 1. The schematic of the DDM. The two boundaries represent the two options in [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Data partitioning example. For an experimental design with three conditions [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The testing data prior and the training data prior. [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Example of posteriors of a single replicate when [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Full posterior mean vs. consensus posterior mean in setting 1. The x-axis [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: RMSE in setting 1. To enhance readability, the RMSE for response bias [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Full posterior SD vs. consensus posterior SD in setting 1. The x-axis indicates [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Example of posteriors for a single replicate when [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Full posterior mean vs. calibrated consensus posterior mean in setting 2. The [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: RMSE in setting 2. To enhance readability, the RMSE for response bias [PITH_FULL_IMAGE:figures/full_fig_p020_10.png]
Figure 11
Figure 11. Figure 11: Full posterior SD vs. calibrated consensus posterior SD in setting 2. The x-axis [PITH_FULL_IMAGE:figures/full_fig_p021_11.png]
Figure 12
Figure 12. Figure 12: Numbers of trials per condition. This figure shows the number of trials per [PITH_FULL_IMAGE:figures/full_fig_p022_12.png]
Figure 13
Figure 13. Figure 13: The pairwise estimation with an imbalanced design. The left panel shows a [PITH_FULL_IMAGE:figures/full_fig_p023_13.png]
Figure 14
Figure 14. Figure 14: Estimated drift rates of 33 brightness strength levels. The blue scatters repres [PITH_FULL_IMAGE:figures/full_fig_p024_14.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

46 extracted references · 29 canonical work pages

  1. [1]

    HDDM: Hierarchical Bayesian estimation of the drift-diffusion model in

    Wiecki, Thomas V and Sofer, Imri and Frank, Michael J , journal=. HDDM: Hierarchical Bayesian estimation of the drift-diffusion model in. 2013 , publisher=

    work page 2013

  2. [2]

    Frontiers in Psychology , volume=

    Systematic parameter reviews in cognitive modeling: Towards a robust and cumulative characterization of psychological processes in the diffusion decision model , author=. Frontiers in Psychology , volume=. 2021 , publisher=

    work page 2021

  3. [3]

    , author=

    A theory of memory retrieval. , author=. Psychological Review , volume=. 1978 , publisher=

    work page 1978

  4. [4]

    arXiv preprint arXiv:2205.01833 , year=

    OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts , author=. arXiv preprint arXiv:2205.01833 , year=

    arXiv

  5. [5]

    Frontiers in Psychology , author =

    A practical introduction to using the drift diffusion model of decision-making in cognitive psychology, neuroscience, and health sciences , volume =. Frontiers in Psychology , author =. 2022 , pages =. doi:10.3389/fpsyg.2022.1039172 , abstract =

    arXiv 2022

  6. [6]

    Frontiers in Computational Neuroscience , volume =

    Neural Substrates of the Drift-Diffusion Model in Brain Disorders , author =. Frontiers in Computational Neuroscience , volume =. 2022 , doi =

    work page 2022

  7. [7]

    Psychonomic Bulletin & Review , volume=

    An EZ-diffusion model for response time and accuracy , author=. Psychonomic Bulletin & Review , volume=. 2007 , publisher=

    work page 2007

  8. [8]

    Psychonomic Bulletin & Review , author =

    Estimating parameters of the diffusion model:. Psychonomic Bulletin & Review , author =. 2002 , pages =. doi:10.3758/BF03196302 , language =

    work page doi:10.3758/bf03196302 2002

Show all 46 references
  1. [9]

    Behavior Research Methods , volume=

    Fast-dm: A free program for efficient diffusion model analysis , author=. Behavior Research Methods , volume=. 2007 , publisher=

    2007

  2. [10]

    Charles , title =

    Geyer, J. Charles , title =. Statistical Science , year =

  3. [11]

    Nature Human Behaviour , author =

    Mental speed is high until age 60 as revealed by analysis of over a million participants , volume =. Nature Human Behaviour , author =. 2022 , pages =. doi:10.1038/s41562-021-01282-7 , language =

    2022 doi

  4. [12]

    Project Implicit , howpublished =

  5. [13]

    Journal of Mathematical Psychology , volume=

    A tutorial on adaptive design optimization , author=. Journal of Mathematical Psychology , volume=. 2013 , publisher=

    2013

  6. [14]

    and Schmitt, Marvin and Schumacher, Lukas and Elsemüller, Lasse and Pratz, Valentin and Schälte, Yannik and Köthe, Ullrich and Bürkner, Paul-Christian , doi =

    Radev, Stefan T. and Schmitt, Marvin and Schumacher, Lukas and Elsemüller, Lasse and Pratz, Valentin and Schälte, Yannik and Köthe, Ullrich and Bürkner, Paul-Christian , doi =. Journal of Open Source Software , month = sep, number =

  7. [15]

    eLife , volume=

    Training deep neural density estimators to identify mechanistic models of neural dynamics , author=. eLife , volume=. 2020 , publisher=

    2020

  8. [16]

    and Mertens, Ulf K

    Radev, Stefan T. and Mertens, Ulf K. and Voss, Andreass and Ardizzone, Lynton and Köthe, Ullrich , year =. doi:10.48550/ARXIV.2003.06281 , abstract =

    2003 doi arXiv

  9. [17]

    Statistica Sinica , pages=

    An overview of composite likelihood methods , author=. Statistica Sinica , pages=. 2011 , publisher=

    2011

  10. [18]

    Annual Review of Statistics and Its Application , volume=

    Neural methods for amortized inference , author=. Annual Review of Statistics and Its Application , volume=. 2024 , publisher=

    2024

  11. [19]

    Frontiers in Psychology , volume=

    Comparing eight parameter estimation methods for the Ratcliff diffusion model using free software , author=. Frontiers in Psychology , volume=. 2020 , publisher=

    2020

  12. [20]

    Behavior Research Methods , volume=

    Extending JAGS: A tutorial on adding custom distributions to JAGS (with a diffusion model example) , author=. Behavior Research Methods , volume=. 2014 , publisher=

    2014

  13. [21]

    Neural Networks , volume=

    Multilayer feedforward networks are universal approximators , author=. Neural Networks , volume=. 1989 , publisher=

    1989

  14. [22]

    Neural Networks , volume=

    On the approximate realization of continuous mappings by neural networks , author=. Neural Networks , volume=. 1989 , publisher=

    1989

  15. [23]

    IEEE Transactions on Neural Networks and Learning Systems , volume=

    BayesFlow: Learning complex stochastic models with invertible neural networks , author=. IEEE Transactions on Neural Networks and Learning Systems , volume=. 2020 , publisher=

    2020

  16. [24]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Normalizing flows: An introduction and review of current methods , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2020 , publisher=

    2020

  17. [25]

    2022 , booktitle=

    Flow Matching for Generative Modeling , author=. 2022 , booktitle=

    2022

  18. [26]

    ACM Computing Surveys , volume=

    Diffusion models: A comprehensive survey of methods and applications , author=. ACM Computing Surveys , volume=. 2023 , publisher=

    2023

  19. [27]

    The Annals of Mathematical Statistics , author =

    On. The Annals of Mathematical Statistics , author =. 1951 , pages =. doi:10.1214/aoms/1177729694 , language =

    1951

  20. [28]

    Dagm german conference on pattern recognition , pages=

    Detecting model misspecification in amortized Bayesian inference with neural networks , author=. Dagm german conference on pattern recognition , pages=. 2023 , organization=

    2023

  21. [29]

    Statistical Science , author =

    Inference from. Statistical Science , author =. 1992 , file =. doi:10.1214/ss/1177011136 , number =

    1992

  22. [30]

    Psychological Science , author =

    Modeling. Psychological Science , author =. 1998 , pages =. doi:10.1111/1467-9280.00067 , abstract =

    1998

  23. [31]

    Reanalysis of

    Singmann, Henrik , year =. Reanalysis of

  24. [32]

    2022 , note =

    rtdists: Response Time Distributions , author =. 2022 , note =

    2022

  25. [33]

    Rank-normalization, Folding, and Localization: An Improved

    Vehtari, Aki and Gelman, Andrew and Simpson, Daniel and Carpenter, Bob and B\"urkner, Paul-Christian , journal =. Rank-normalization, Folding, and Localization: An Improved. 2021 , doi =

    2021

  26. [34]

    arXiv preprint arXiv:2602.07098 , year=

    BayesFlow 2.0: Multi-Backend Amortized Bayesian Inference in Python , author=. arXiv preprint arXiv:2602.07098 , year=

  27. [35]

    Big Data and Information Theory , pages=

    Bayes and big data: The consensus Monte Carlo algorithm , author=. Big Data and Information Theory , pages=. 2022 , publisher=

    2022

  28. [36]

    The Annals of Mathematical Statistics , volume=

    An optimum property of regular maximum likelihood estimation , author=. The Annals of Mathematical Statistics , volume=. 1960 , publisher=

    1960

  29. [37]

    Pseudolikelihood estimation: some examples , author=. Sankhy. 1991 , publisher=

    1991

  30. [38]

    Statistica Sinica , pages=

    Bayesian inference from composite likelihoods, with an application to spatial extremes , author=. Statistica Sinica , pages=. 2012 , publisher=

    2012

  31. [39]

    Statistica Sinica , pages=

    Bayesian composite marginal likelihoods , author=. Statistica Sinica , pages=. 2011 , publisher=

    2011

  32. [40]

    Composite likelihood methods , author=. Statistical Inference from Stochastic Processes: Proceedings of the AMS-IMS-SIAM Joint Summer Research Conference Held August 9-15, 1987, with Support from the National Science Foundation and the Army Research Office , volume=. 1988 , or...

    1987

  33. [41]

    2005 , publisher=

    Models for discrete longitudinal data , author=. 2005 , publisher=

    2005

  34. [42]

    Wiley Interdisciplinary Reviews: Computational Statistics , volume=

    Importance sampling: a review , author=. Wiley Interdisciplinary Reviews: Computational Statistics , volume=. 2010 , publisher=

    2010

  35. [43]

    arXiv preprint arXiv:2411.12068 , year=

    The statistical accuracy of neural posterior and likelihood estimation , author=. arXiv preprint arXiv:2411.12068 , year=

    arXiv

  36. [44]

    2012 , publisher=

    Asymptotic methods in statistical decision theory , author=. 2012 , publisher=

    2012

  37. [45]

    arXiv preprint arXiv:1311.4780 , year=

    Asymptotically exact, embarrassingly parallel MCMC , author=. arXiv preprint arXiv:1311.4780 , year=

    arXiv

  38. [46]

    Statistics and Computing , volume=

    Graphical test for discrete uniformity and its applications in goodness-of-fit evaluation and multiple sample comparison , author=. Statistics and Computing , volume=. 2022 , publisher=

    2022

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.