JADAI: Jointly Amortizing Adaptive Design and Bayesian Inference

Desi R. Ivanova; Lars K\"uhmichel; Niels Bracher; Paul-Christian B\"urkner; Stefan T. Radev; Xavier Intes

REVIEW 2 major objections 5 minor 1 cited by

JADAI: Jointly Amortizing Adaptive Design and Bayesian Inference

T0 review · 2 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read JADAI shows that adaptive experimental design and Bayesian posterior inference can be trained together, end to end, around a single objective that rewards reducing posterior error at every step of an experiment.

desk verdict A well-engineered joint policy+inference framework with strong empirical results, whose main theoretical hinge for diffusion models is an unproven proxy — worth serious refereeing, not a desk reject. read the letter →

arxiv 2512.22999 v2 pith:YJHUBZ4A submitted 2025-12-28 stat.ML cs.AIcs.LG

Niels Bracher , Lars Kühmichel , Desi R. Ivanova , Xavier Intes , Paul-Christian Bürkner , Stefan T. Radev This is my paper · ORCID

classification stat.MLcs.AIcs.LG

keywords Bayesianexperimentaldesignadaptiveamortizedinferencesimulation-baseddiffusionmodelsflowmatchingposteriorestimationpolicy

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

Bayesian adaptive design (choosing each next measurement to maximize information) and posterior inference (estimating parameters from the measurements) have usually been solved as separate problems: a design policy proposes experiments, and a slow inference method is run afterwards. This paper tries to establish that the two can be amortized jointly: one policy network chooses each next design, one history network summarizes the measurements so far, and one diffusion- or flow-matching network estimates the posterior at every step, all trained end-to-end on a scalar utility that sums per-step reductions in posterior error. The claim is that maximizing this utility, even though it is not an exact expected-information-gain objective when the posterior is an implicit generative model, yields policies competitive with or better than existing approaches on standard adaptive-design benchmarks, including a high-dimensional multimodal image-discovery task. If correct, this turns sequential experiment design and inference into a single fast, fully amortized step rather than a two-stage loop.

What carries the argument

The load-bearing object is the incremental posterior-loss utility u_T(θ, h_{0:T}) = Σ_{t=0}^T (ℓ*_{t-1} - ℓ_t), with ℓ_{-1}=0 and ℓ*_t = detach(ℓ_t), where ℓ_t is the per-step posterior estimation loss conditioned on the history summary. For a normalized density model this sum telescopes exactly and reproduces the lower bound on total expected information gain; for implicit models it is used as a proxy. A second mechanism is the detached-history design pass: the policy receives a summary with gradients stopped, so the history network is trained only through the posterior losses while the policy is trained through the observations its designs produce; this removes nested backpropagation throu

What would settle it

In a small Gaussian adaptive-design problem where the true expected information gain is computable in closed form, train JADAI with a deliberately miscalibrated posterior estimator so that posterior loss decreases while the posterior is inaccurate, then compare the resulting policy's estimated information gain or final inference accuracy against a policy trained on the true objective; a measurable gap would show the surrogate is not tracking information gain.

Watch

Extended reading notes

Core claim

The paper's central discovery is an objective identity: for a normalized-density posterior estimator, the variational lower bound on total expected information gain telescopes into a sum over experiment steps of per-step differences in posterior loss, log q(θ|h_t) - log q(θ|h_{t-1}). JADAI builds on this by using the same sum, u_T = Σ_{t=0}^T (ℓ*_{t-1} - ℓ_t), as a generic training utility even when the posterior estimator is an implicit diffusion or flow-matching model whose log-density is unavailable, with the previous loss value detached. Although the variational bound interpretation no longer applies, the objective still drives rollouts along which posterior loss decreases at every step.

Load-bearing premise

The claim rests on the detached per-step posterior-loss difference (ℓ_{t-1} - ℓ_t) being a faithful proxy for expected information gain when the posterior is an implicit diffusion or flow-matching model; if that proxy misaligns with true information gain, the policy can be optimal for the surrogate while poor for the actual experiment.

Editorial extensions

If this is right

At any point in an experiment, JADAI provides an approximate posterior over parameters, not only after the full horizon, enabling active data-acquisition workflows.
A policy trained to minimize posterior losses over a longer horizon remains competitive at shorter horizons, so longer training horizons can improve early decisions.
Diffusion- and flow-matching-based posterior estimators can be inserted into the same design objective, making high-dimensional multimodal posteriors tractable in adaptive design.
When the design policy is replaced by a random policy, performance falls to baseline levels, indicating that most of the gain comes from the learned policy rather than the posterior network.
Test-time rollouts take milliseconds on low-dimensional tasks and are roughly an order of magnitude faster than prior methods on the high-dimensional benchmark.

Reading between the lines

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

Beyond the paper: if the per-step loss proxy is as reliable as the experiments suggest, the same utility should extend to other implicit posterior families, such as consistency models or adversarial generators, without re-deriving an information-theoretic objective.
Beyond the paper: because the utility is defined entirely through posterior losses, JADAI could be coupled with gradient-free design optimization or policy-gradient estimators to handle black-box simulators; the paper only tests differentiable simulators.
Beyond the paper: the paper notes but does not implement sharing a transformer backbone across policy and history networks; doing so could improve sample efficiency on harder design spaces.
Beyond the paper: a testable consequence is that deliberately miscalibrating the posterior estimator, while keeping the posterior loss artificially low, should degrade the learned designs; monitoring this gap would validate the surrogate.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The paper does something genuinely useful: it trains a policy, a history encoder, and a diffusion/flow-matching posterior estimator end-to-end for adaptive design, using the detach-and-telescope trick to turn per-step posterior losses into a training signal that also shapes the policy. That is a clean, practical idea, and the experiments back it up: the method beats or matches strong baselines on Location Finding and CES, and on MNIST Image Discovery it is clearly ahead of CoDiff. The ablation study is honest — nested BPTT is unstable, truncated BPTT hurts, fixed rollout length helps terminal but not intermediate horizons. That kind of reporting earns credit.

The soft spot is exactly where the stress test lands. For a normalized density estimator, Eq. (14) is a real telescoping of the Barber–Agakov bound, so the objective has a clean information-theoretic reading. For diffusion and flow matching, ℓ_t is a regression loss, and the difference ℓ_{t-1} − ℓ_t is not a log-density ratio, not a bound on incremental EIG, and not an unbiased estimator of the score difference in Eq. (15). The paper says so itself in Section 3.2 — 'the Barber–Agakov interpretation is no longer the same' — and then appeals to experiments. That is an honest concession, but it leaves the central claim resting entirely on empirical performance. If a diffusion model is miscalibrated, the policy could reduce score-matching loss along easier trajectories without gaining information. No calibration check or monotonicity argument is offered. This is a real gap, but I would not call it fatal: the objective is a plausible surrogate, the sPCE and SSIM evaluations are external to the training loss, and the results are strong enough to be interesting even without a formal guarantee. The differentiability requirement is also acknowledged, and it is a genuine limitation for black-box simulators.

The citation pattern looks fine. The paper builds on the relevant DAD/iDAD/ALINE line and cites it properly. Some self-citation appears in the diffusion tutorial reference, but it is not load-bearing.

Who is this for? People working on amortized Bayesian experimental design or simulation-based inference with diffusion posteriors. It deserves a serious referee: the empirical contribution is substantial, the method is clearly described, and the missing theoretical link is the thing a reviewer should push on. I would engage with it and would want the authors to either provide a formal argument for the proxy or add calibration diagnostics showing that the diffusion posterior tracks the true posterior along policy rollouts.

Referee Report

2 major / 5 minor

Summary. The paper introduces JADAI, a framework that jointly amortizes Bayesian adaptive design and posterior inference by training a design policy, a history summary network, and a generative posterior estimator end-to-end. The training objective is the generic utility u_T = Σ_{t=0}^T (ℓ*_{t-1} − ℓ_t), where ℓ_t is a per-step posterior loss and ℓ* denotes a detached copy. For a normalized density estimator (e.g., a normalizing flow), this utility is a telescoping form of the Barber–Agakov lower bound on total expected information gain (Eqs. 12–14). For diffusion and flow-matching estimators, ℓ_t is instead a weighted score-matching or flow-matching regression loss (Eqs. 27 and 34), and the paper explicitly states that the Barber–Agakov interpretation no longer applies and appeals to experimental performance. Experiments on Location Finding, CES, and MNIST Image Discovery report competitive or superior sPCE and SSIM/NRMSE relative to recent baselines, with ablations in Appendix B.

Significance. If the empirical findings hold, JADAI is a practically valuable contribution: it extends amortized adaptive design to high-dimensional, multimodal posteriors without reinforcement-learning value networks, and it provides posterior estimates at every acquisition step. The normalized-density case gives a clean connection to TEIG, and the experiments are substantial, including random-policy controls that support the claim that the performance gap in Image Discovery comes primarily from the learned policy (Section 5.3, Figure 6). The paper is also transparent about limitations, including the heuristic status of the implicit-model objective (Section 3.2) and the need for differentiable simulators (Section 6). The main weakness is theoretical: for the diffusion/flow setting that powers the flagship high-dimensional results, the objective is not shown to track or bound expected information gain, so the paper's formal claim is not established for its primary regime.

major comments (2)

[Section 3.2, Eqs. (16), (27), (34)] The central claim that JADAI maximizes information gain rests on treating ℓ_{t-1}−ℓ_t as incremental information. For a normalized density estimator, Eq. (14) is an identity and the Barber–Agakov bound is clean. For diffusion models (Eq. 27) and flow matching (Eq. 34), ℓ_t is a regression loss against noise/velocity targets; ℓ_{t-1}−ℓ_t is not a log-density ratio and is not shown to bound, estimate, or track incremental EIG. The manuscript concedes this in Section 3.2 and appeals to experiments. This is load-bearing because the policy is trained through this difference. A policy can reduce ℓ_t by steering to histories that are easy for the current posterior estimator without increasing true information gain unless additional calibration or monotonicity holds. I ask for either (a) a formal statement of conditions under which ℓ_{t-1}−ℓ_t is a valid surrogate for incremental information, or
[Section 5.3, Table 2, Figure 4] For the high-dimensional Image Discovery benchmark, the central evaluation is SSIM/NRMSE against the ground-truth image, not an information-theoretic measure of posterior quality. The random-policy controls in Figure 6 are informative, but SSIM is a perceptual reconstruction metric and can be high even if the posterior is miscalibrated. To support the claim that the learned policy improves inference rather than only reconstruction, please report a posterior calibration or coverage diagnostic at each measurement step, for example simulation-based calibration or conditional coverage on a validation set.

minor comments (5)

[Section 3.2, Eq. (15)] The notation suggests a score with respect to θ_τ, but in the diffusion derivation the score is with respect to the latent z_τ; please align notation with Eqs. (23)–(24).
[Section 5.2 / Appendix C.3] The heading 'Constant Elasticity of Substitutions' should be 'Constant Elasticity of Substitution' for consistency with the CES benchmark name.
[Table 1] Entries marked '−' are not defined in the caption; please specify whether they indicate 'not reported' or 'not run'.
[Section 5.1, Figure 2] The corner plot axis labels are difficult to read at the current resolution; please ensure they are legible in the final version.
[Section 6] The requirement of a differentiable simulator is stated clearly, but it should also be mentioned in the abstract or introduction, since it materially constrains the scope of the method.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the JADAI objective is a stated heuristic surrogate, not a hidden re-use of its evaluation targets, and the self-citations are background only.

full rationale

The paper's central derivation replaces the exact Barber–Agakov TEIG lower bound, which telescopes only for normalized density estimators (Eqs. 12–14), with the generic loss-difference utility u_T = sum(ℓ*_{t-1} − ℓ_t) for diffusion and flow-matching posterior estimators. The authors explicitly acknowledge that for these implicit models “the Barber–Agakov interpretation is no longer the same” (Section 3.2) and justify the surrogate empirically rather than theoretically. This is a heuristic modeling choice and a limitation, not circular reasoning: the training objective is not defined in terms of the evaluation metrics (sPCE, SSIM, NRMSE), and those metrics are computed externally from the simulator and ground-truth images. The telescoping observation that u_T forward-equals −ℓ_T is explicitly handled by gradient detachment so the method intentionally optimizes all intermediate losses rather than only the final one; this is an algorithmic design, not a circular reduction. The self-citations (e.g., Arruda et al. 2025, Radev et al. 2020, Schmitt et al. 2024) are used for background on diffusion-based SBI and summary networks, not as load-bearing evidence for the paper's new claims. No fitted parameter is relabeled as a prediction, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. The derivation chain is therefore self-contained with respect to circularity, and the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

No new physical entities or mediators are introduced. The framework uses standard neural-network components; the free parameters listed are training schedules and noise schedules that are hand-chosen but not fitted to the target result. The main axiomatic burden is the heuristic posterior-loss proxy for information gain in implicit models and the differentiability requirement.

free parameters (3)

Exploration schedule ρ_n = Cosine decay from 1 to 0 over first 30% of ID training; ρ=0 for LF/CES after pretraining
Hand-chosen curriculum that mixes random designs with policy outputs during training. It affects the diversity of observed data but is not a fitted constant of the model; the central claim does not hinge on the exact schedule.
Rollout length schedule R(n) = Linear from 2 to 6 for ID; fixed T for LF/CES
Curriculum on the maximum rollout length to stabilize training. Standard practice; not a parameter that changes the method's theoretical grounding.
Diffusion noise schedule = Cosine schedule λ(τ) = -2 log(tan(πτ/2))
A standard noise schedule chosen by hand in diffusion models; affects loss weighting but is not fitted to data and does not determine the core methodology.

assumptions (3)

domain assumption The simulator is differentiable with respect to design variables and random states can be reparameterized.
The gradient-based training requires backpropagating through the simulator. The paper states in the Conclusion: 'our experiments rely on differentiating through the simulator.' Many real simulators (e.g., discrete-event, agent-based) are non-differentiable, limiting applicability.
ad hoc to paper The per-step posterior loss ℓ_t is a monotonically valid proxy for information gained about θ.
For implicit generative models (diffusion/flow), no theoretical link between ℓ_{t-1}-ℓ_t and expected information gain is provided. The method's effectiveness rests on this heuristic, as acknowledged in Section 3.2.
domain assumption The summary network η_ω produces approximately sufficient statistics even when the history is detached before the policy.
Section 3.4 detaches the history state before the policy (ξ_t = π_φ(h*_{t-1})) to avoid nested BPTT. This assumes the summary remains informative for design choices without gradient shaping from the policy, which may fail in more complex design spaces.

how reviews work

0 comments

Cite this review

Pith. "Pith review of JADAI: Jointly Amortizing Adaptive Design and Bayesian Inference." pith.science (2026). https://pith.science/paper/YJHUBZ4A

@misc{pith2026251222999,
  author       = {Pith},
  title        = {Pith review of: JADAI: Jointly Amortizing Adaptive Design and Bayesian Inference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YJHUBZ4A}},
  note         = {Machine review of arXiv:2512.22999}
}

read the original abstract

We consider problems of parameter estimation where design variables can be actively optimized to maximize information gain. To this end, we introduce JADAI, a framework that jointly amortizes Bayesian adaptive design and inference by training a policy, a history network, and an inference network end-to-end. The networks minimize a generic loss that aggregates incremental reductions in posterior error along experimental sequences. Inference networks are instantiated with diffusion-based posterior estimators that can approximate high-dimensional and multimodal posteriors at every experimental step. Across standard adaptive design benchmarks, JADAI achieves superior or competitive performance.

Figures

Figures reproduced from arXiv: 2512.22999 by the authors.

**Figure 1.** Overview of amortized SBI, BAD, and our proposed JADAI framework. Left: Amortized SBI, where a neural posterior estimator qψ(θ | x) is trained on simulator pairs (x, θ) under a fixed design ξ. Middle: Amortized BAD, where a designer runs a full T-step rollout for a fixed parameter θ: a policy πϕ maps the history state ht−1 = ηω({(ξk, xk)} t−1 k=1) to a new design ξt, and the simulator returns xt, iteratively forming… view at source ↗

**Figure 2.** Rollout process for Location Finding. Panels: posterior samples and chosen designs over time t, with crosses marking the true source locations. The second posterior mode is typically uncovered around t = 10 measurements. Bottom right: corner plot of the learned posterior over the two sources at t = 10 shows nearly identical densities at (θ11, θ12) and (θ21, θ22), indicating that the model correctly captures exchange… view at source ↗

**Figure 3.** Image discovery rollout. Each column is one measurement step. From top to bottom: ground-truth digit, cumulative measurements with the newly chosen design highlighted in blue, one posterior sample, and the posterior mean over 100 samples. The correct digit class is typically identified after 1–2 measurements, after which the policy mainly refines local structure in uncertain regions [PITH_FULL_IMAGE:figures/full_… view at source ↗

Figures from the paper (4 more)

**Figure 5.** Figure 5: Additional rollout examples for location finding (DAD setting, cf [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

**Figure 6.** Figure 6: Random policy validation SSIM (↑) and NRMSE (↓) as a function of the number of measurements. Shaded bands indicate the interquartile range over the validation set. Using the random policy, our posterior samples do not differ significantly from CoDiff’s in SSIM, indicat…

**Figure 7.** Figure 7: Additional validation samples for the MNIST image discovery experiment with policy, using a diffusion model with σ = 0.001. Each block shows one rollout process, where the columns represent the measurement steps. The first row shows the ground truth image. The second r…

**Figure 8.** Figure 8: Difference in log-scale signal support for noise-free and noisy measurements in image discovery. Left: The noise-free measurement clearly shows the full digit shape in just one measurement, even outside of the intended measurement mask. Right: Using additive noise prev…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Bayesian Experimental Design via Score Matching
stat.ML 2026-07 conditional novelty 7.0 of 10

SCOREBED isolates EIG double intractability in a policy-independent score-matching stage, then trains design policies with a singly intractable gradient estimator, enabling cheap multi-policy selection.

Reference graph

Works this paper leans on

14 extracted references · 2 canonical work pages · cited by 1 Pith paper

[11]

and Li, Q

5 Wang, Z. and Li, Q. Information content weighting for perceptual image quality assessment.IEEE Transactions on Image Processing, 20(5):1185–1198, 2011. doi: 10.1 109/TIP.2010.2092435. 16 Wang, Z., Bovik, A., Sheikh, H., and Simoncelli, E. Image quality assessment: from error visibility to structural similarity.IEEE Transactions on Image Processing, 13 (...

arXiv 2011
[48]

and Gutmann, M

13 Kleinegesse, S. and Gutmann, M. U. Bayesian experimental design for implicit models by mutual information neural estimation.arXiv [stat.ML], February 2020. URL http: //arxiv.org/abs/2002.08129. 5 Kuck, H., de Freitas, N., and Doucet, A. Smc samplers for bayesian optimal nonlinear design. In2006 IEEE Nonlinear Statistical Signal Processing Workshop, pp....

arXiv 2020
[72]

1 Rainforth, T., Foster, A., Ivanova, D

URL http://dx.doi.org/10.1371/journal.pcbi.1009472. 1 Rainforth, T., Foster, A., Ivanova, D. R., and Bickford Smith, F. Modern bayesian experimental design.Statistical Science, 39(1):100–114, 2024. 2, 3 Rezende, D. and Mohamed, S. Variational inference with normalizing flows. InInternational conference on ma- chine learning, pp. 1530–1538. PMLR, 2015. 2 S...

work page doi:10.1371/journal.pcbi.1009472 2024
[685]

and Agakov, F

1, 5, 13, 14 Barber, D. and Agakov, F. The IM algorithm: a variational approach to Information Maximization.Neural Informa- tion Processing Systems, pp. 201–208, December 2003. URL https://papers.nips.cc/paper files/paper/2003/file/ a6ea8471c120fe8cc35a2954c9b9c595-Paper.pdf. 3, 4, 5 Blau, T., Bonilla, E. V ., Chades, I., and Dezfouli, A. Op- timizing seq...

arXiv 2003
[924]

Method Details A.1

8 12 JADAI: Joint Adaptive Design and Amortized Inference A. Method Details A.1. Diffusion model background In the following, we provide a brief overview of our diffusion model (Ho et al., 2020; Song & Ermon, 2019; Kingma & Gao,

2020
[1093]

and Marzouk, Y

22 Huan, X. and Marzouk, Y . M. Simulation-based optimal bayesian experimental design for nonlinear systems.Jour- nal of Computational Physics, 232(1):288–317, 2013. 5 Huang, D., Guo, Y ., Acerbi, L., and Kaski, S. Amortized bayesian experimental design for decision-making.Ad- vances in Neural Information Processing Systems, 37: 109460–109486, 2024. 6, 8 ...

arXiv 2013
[2001]

Deep residual learn- ing for image recognition

5 He, K., Zhang, X., Ren, S., and Sun, J. Deep residual learn- ing for image recognition. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2016. 22 Hedman, M., Ivanova, D. R., Guan, C., and Rainforth, T. Step-dad: Semi-amortized policy-based bayesian exper- imental design.International conference on machine lear...

arXiv 2016
[2015]

doi: https://doi.org/10.1016/j.ae ue.2014.09.002

ISSN 1434-8411. doi: https://doi.org/10.1016/j.ae ue.2014.09.002. URL https://www.sciencedirect.com/sc ience/article/pii/S1434841114002519. 16 Hahn, C., Lemos, P., Parker, L., R ´egaldo-Saint Blancard, B., Eickenberg, M., Ho, S., Hou, J., Massara, E., Modi, C., Moradinezhad Dizgah, A., and Spergel, D. Cos- mological constraints from non-gaussian and nonli...

work page doi:10.1016/j.ae 2014

Show all 14 references

[2018]

21 Song, Y

URL https://arxiv.org/abs/1708.07120. 21 Song, Y . and Ermon, S. Generative Modeling by Estimating Gradients of the Data Distribution.arXiv [cs.LG], July

arXiv
[2019]

2, 13 Søgaard, J., Krasula, L., Shahid, M., Temel, D., Brunnstr¨om, K., and Razaak, M

URL http://arxiv.org/abs/1907.05600. 2, 13 Søgaard, J., Krasula, L., Shahid, M., Temel, D., Brunnstr¨om, K., and Razaak, M. Applicability of existing objective metrics of perceptual quality for adaptive video streaming. volume 2016, 02 2016. doi: 10.2352/ISSN.2470-1173. 2016.1...

1907 arXiv
[2023]

r1GL/talcOUHCzu7TmEknpH/Tl8=

used to approximate the posterior p(θ| ·)≈q ψ(θ| ·). More details on diffusion models in a general setting can be found in (Karras et al., 2022) and, for simulation-based inference, in (Arruda et al., 2025). A diffusion model learns how to gradually denoise a sample from a bas...

2022
[2024]

1, 5 Sharrock, L., Simons, J., Liu, S., and Beaumont, M

URL https://openreview.net/forum?id=IHjKpKlj yH. 1, 5 Sharrock, L., Simons, J., Liu, S., and Beaumont, M. Sequen- tial neural score estimation: Likelihood-free inference with conditional score based diffusion models.arXiv [stat.ML], October 2022. URL http://arxiv.org/abs/2210 ...

2022
[2025]

doi: 10.1146/annure v-statistics-112723-034123

ISSN 2326-8298,2326-831X. doi: 10.1146/annure v-statistics-112723-034123. URL http://dx.doi.org/10.11 46/annurev-statistics-112723-034123. 1 Zhang, X., Huang, D., Kaski, S., and Martinelli, J. PABBO: Preferential amortized black-box optimization.arXiv [stat.ML], March 2025. UR...

2025 doi
[2138]

URL http://dx.doi.org/10.1016/j.cma.2024.117457

doi: 10.1016/j.cma.2024.117457. URL http://dx.doi.org/10.1016/j.cma.2024.117457. 5 Drovandi, C. C., McGree, J. M., and Pettitt, A. N. A se- quential monte carlo algorithm to incorporate model un- certainty in bayesian sequential design.Journal of Com- putational and Graphical ...

2024

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [11]

and Li, Q

5 Wang, Z. and Li, Q. Information content weighting for perceptual image quality assessment.IEEE Transactions on Image Processing, 20(5):1185–1198, 2011. doi: 10.1 109/TIP.2010.2092435. 16 Wang, Z., Bovik, A., Sheikh, H., and Simoncelli, E. Image quality assessment: from error visibility to structural similarity.IEEE Transactions on Image Processing, 13 (...

arXiv 2011

[2] [48]

and Gutmann, M

13 Kleinegesse, S. and Gutmann, M. U. Bayesian experimental design for implicit models by mutual information neural estimation.arXiv [stat.ML], February 2020. URL http: //arxiv.org/abs/2002.08129. 5 Kuck, H., de Freitas, N., and Doucet, A. Smc samplers for bayesian optimal nonlinear design. In2006 IEEE Nonlinear Statistical Signal Processing Workshop, pp....

arXiv 2020

[3] [72]

1 Rainforth, T., Foster, A., Ivanova, D

URL http://dx.doi.org/10.1371/journal.pcbi.1009472. 1 Rainforth, T., Foster, A., Ivanova, D. R., and Bickford Smith, F. Modern bayesian experimental design.Statistical Science, 39(1):100–114, 2024. 2, 3 Rezende, D. and Mohamed, S. Variational inference with normalizing flows. InInternational conference on ma- chine learning, pp. 1530–1538. PMLR, 2015. 2 S...

work page doi:10.1371/journal.pcbi.1009472 2024

[4] [685]

and Agakov, F

1, 5, 13, 14 Barber, D. and Agakov, F. The IM algorithm: a variational approach to Information Maximization.Neural Informa- tion Processing Systems, pp. 201–208, December 2003. URL https://papers.nips.cc/paper files/paper/2003/file/ a6ea8471c120fe8cc35a2954c9b9c595-Paper.pdf. 3, 4, 5 Blau, T., Bonilla, E. V ., Chades, I., and Dezfouli, A. Op- timizing seq...

arXiv 2003

[5] [924]

Method Details A.1

8 12 JADAI: Joint Adaptive Design and Amortized Inference A. Method Details A.1. Diffusion model background In the following, we provide a brief overview of our diffusion model (Ho et al., 2020; Song & Ermon, 2019; Kingma & Gao,

2020

[6] [1093]

and Marzouk, Y

22 Huan, X. and Marzouk, Y . M. Simulation-based optimal bayesian experimental design for nonlinear systems.Jour- nal of Computational Physics, 232(1):288–317, 2013. 5 Huang, D., Guo, Y ., Acerbi, L., and Kaski, S. Amortized bayesian experimental design for decision-making.Ad- vances in Neural Information Processing Systems, 37: 109460–109486, 2024. 6, 8 ...

arXiv 2013

[7] [2001]

Deep residual learn- ing for image recognition

5 He, K., Zhang, X., Ren, S., and Sun, J. Deep residual learn- ing for image recognition. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2016. 22 Hedman, M., Ivanova, D. R., Guan, C., and Rainforth, T. Step-dad: Semi-amortized policy-based bayesian exper- imental design.International conference on machine lear...

arXiv 2016

[8] [2015]

doi: https://doi.org/10.1016/j.ae ue.2014.09.002

ISSN 1434-8411. doi: https://doi.org/10.1016/j.ae ue.2014.09.002. URL https://www.sciencedirect.com/sc ience/article/pii/S1434841114002519. 16 Hahn, C., Lemos, P., Parker, L., R ´egaldo-Saint Blancard, B., Eickenberg, M., Ho, S., Hou, J., Massara, E., Modi, C., Moradinezhad Dizgah, A., and Spergel, D. Cos- mological constraints from non-gaussian and nonli...

work page doi:10.1016/j.ae 2014