Pith. sign in

REVIEW 3 minor 2 references

Diffusion model priors in a Bayesian setup reconstruct rainfall fields from microwave link data more accurately than standard baselines.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.3

2026-06-30 23:36 UTC pith:7S3JU73E

load-bearing objection The paper frames CML rainfall reconstruction as a Bayesian inverse problem with diffusion model priors and reports gains over Gaussian process baselines on synthetic and real data.

arxiv 2605.05520 v2 pith:7S3JU73E submitted 2026-05-06 cs.LG stat.APstat.ML

Bayesian Rain Field Reconstruction using Commercial Microwave Links and Diffusion Model Priors

Badr Moufad , Albina Ilina , Hai Victor Habi , Salem Lahlou , Yazid Janati , Hagit Messer , Eric Moulines This is my paper
classification cs.LG stat.APstat.ML
keywords rainfall reconstructioncommercial microwave linksdiffusion modelsBayesian inverse problemsspatial priorsposterior samplingline integration
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper frames rain field reconstruction from path-integrated commercial microwave link measurements as a Bayesian inverse problem. It uses pre-trained diffusion models as spatial priors to enable training-free posterior sampling through methods such as Plug-and-Play and Sequential Monte Carlo. This yields reconstructions that better preserve key rainfall statistics than censored Gaussian process priors. Experiments on synthetic and real datasets show consistent gains over established CML reconstruction approaches that oversimplify the line integration.

Core claim

Diffusion models serve as high-fidelity spatial priors in the Bayesian inverse problem of recovering ground-level rain fields from line-integrated attenuation measurements, outperforming censored Gaussian processes in preserving rainfall statistics while supporting a range of sampling algorithms without any model retraining.

What carries the argument

Bayesian inverse problem with diffusion model priors on spatial rain fields, which supports training-free posterior sampling to invert path-integrated CML observations.

Load-bearing premise

Pre-trained diffusion models supply accurate spatial priors for rainfall fields without needing any domain-specific retraining or adaptation.

What would settle it

On a held-out real CML dataset, the diffusion-prior reconstructions fail to match or exceed the rainfall statistics achieved by a censored Gaussian process prior under identical line-integration modeling.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The approach handles heterogeneous precipitation better by respecting the actual line-integration physics rather than treating links as point sensors.
  • A family of sampling methods becomes available for the same prior without additional training.
  • Reconstructed fields exhibit improved fidelity to observed rainfall distributions on both synthetic and real data.
  • The method produces consistent gains over prior CML baselines that neglect line integration.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same diffusion-prior Bayesian framing could apply to other path-integrated sensing problems such as tomography or atmospheric tomography.
  • Efficient sampling variants might enable near-real-time rain mapping if computational cost is reduced further.
  • Combining the diffusion prior with additional sensor modalities could further constrain the inverse problem.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 3 minor

Summary. The manuscript frames commercial microwave link (CML) rain-field reconstruction as a Bayesian inverse problem that employs pre-trained diffusion models as spatial priors. It reports that diffusion models preserve key rainfall statistics more faithfully than censored Gaussian processes, enables training-free posterior sampling via Plug-and-Play, Sequential Monte Carlo, and Replica Exchange samplers, and obtains consistent gains over established CML baselines on both synthetic and real-world data.

Significance. If the empirical gains hold under the stated forward model, the work supplies a practical route for injecting high-capacity generative priors into geophysical inverse problems without domain-specific fine-tuning of the prior. The explicit inclusion of the line-integration likelihood and the training-free sampling strategy are concrete strengths that distinguish the contribution from purely data-driven regression approaches.

minor comments (3)
  1. [Abstract] Abstract: the statement that diffusion models 'better preserve key rainfall statistics' should be accompanied by the specific metrics (e.g., power-spectrum error, wet-area ratio, or exceedance probabilities) and the exact comparison protocol used against censored GPs.
  2. The manuscript should clarify whether the pre-trained diffusion models were used exactly as released or whether any rainfall-specific normalization or conditioning was applied before sampling; this detail affects reproducibility of the 'training-free' claim.
  3. Figure captions and table headers should explicitly state the number of independent realizations, the random-seed protocol, and whether error bars represent standard deviation across realizations or across CML configurations.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, accurate summary of the contributions, and recommendation for minor revision. No major comments were provided for us to address.

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper presents rainfall reconstruction as a Bayesian inverse problem with pre-trained diffusion models serving as spatial priors, followed by training-free posterior sampling via PnP, SMC, and Replica Exchange. All load-bearing steps (likelihood construction via explicit line-integration model, prior comparison to censored GPs, and empirical gains on synthetic/real CML data) are externally grounded in pre-trained models and dataset experiments rather than reducing to fitted parameters, self-definitions, or self-citation chains by construction. No equations or claims in the provided text exhibit the enumerated circular patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities; the domain assumption that diffusion models act as suitable rainfall priors is implicit but not quantified.

axioms (1)
  • domain assumption Diffusion models trained on general data serve as high-fidelity spatial priors for rainfall fields.
    Invoked when claiming better preservation of rainfall statistics than Gaussian processes.

pith-pipeline@v0.9.1-grok · 5685 in / 1084 out tokens · 23626 ms · 2026-06-30T23:36:53.101602+00:00 · methodology

0 comments
read the original abstract

Commercial Microwave Links (CMLs) offer dense spatial coverage for rainfall sensing but produce path-integrated measurements that make accurate ground-level reconstruction challenging. Existing methods typically oversimplify CMLs as point sensors and neglect line integration relating rainfall to signal attenuation, resulting in degraded performance under heterogeneous precipitation. In this work, we view rain field reconstruction as a Bayesian inverse problem with Diffusion Models (DMs) as high-fidelity spatial priors. We show that diffusion models better preserve key rainfall statistics compared to censored Gaussian processes. Framing rainfall estimation as a Bayesian inverse problem with a DM prior enables training-free posterior sampling using a broad family of methods, including Plug-and-Play, Sequential Monte Carlo, and Replica Exchange methods. Experiments on synthetic and real-world datasets demonstrate consistent improvements over established CML-based reconstruction baselines.

Figures

Figures reproduced from arXiv: 2605.05520 by Albina Ilina, Badr Moufad, Eric Moulines, Hagit Messer, Hai Victor Habi, Salem Lahlou, Yazid Janati.

Figure 1
Figure 1. Figure 1: Comparison of rain field statistics between the refer￾ence samples and generated samples using the DM prior. The histograms show densities and are built using 5000 samples. exists a rich literature on stochastic precipitation model￾ing and geostatistical reconstruction (Wilks & Wilby, 1999; Wheater et al., 2000; Yang et al., 2005). A classical and widely used statistical abstraction models this mixed behav… view at source ↗
Figure 2
Figure 2. Figure 2: Examples of generated rain fields using the prior DM compared with reference rain fields. A total of 5,000 samples were generated. In each subfigure, on the left, samples are randomly drawn from the datasets; on the right, they are drawn from the top 90% wet samples by cumulative rain. clean sample x0 in the above transition with the output of a neural network (t, xt) 7→ Dθ t (xt) parameterized by θ. This … view at source ↗
Figure 3
Figure 3. Figure 3: Comparison between the reconstructions of the baselines on the inverse problem with diffusion prior on the setting of GP. The y-axis shows intensities while the x-axis represents a one-dimensional grid of the [−5, 5] interval. The dashed horizontal lines depict values of the integral over the observed intervals. likelihood of the inverse problem is p(y | x0) ∝ exp{−1 2 ∥y − M(x0)∥ 2 Σ−2 }, (5) with ∥v∥ 2 Σ… view at source ↗
Figure 4
Figure 4. Figure 4: Comparisons of rain field reconstructions on real CMLs links from OpenMRG. We depict the network of CMLs in red view at source ↗
Figure 5
Figure 5. Figure 5: Power-law parameters of the CML as a function of link’s length and the frequency on the OpenMRG dataset. C.1. Meteorological baselines IDW. We implement inverse-distance weighting (IDW) following Eshel et al. (2021); Shepard (1968). Each CML link observation is represented as a point measurement located at the link midpoint. For every grid location s, we compute a normalized weighted average using p = 2 an… view at source ↗
Figure 6
Figure 6. Figure 6: Illustration of a line sensor on a 6 × 4 grid. The blue dots define the line sensor whereas the red dots mark its intersections with the two-dimensional grid. The color intensity of each cell is proportional to the intersection length. Considered convention. In our setting, the radar maps use the origin (− 1 2 , − 1 2 ) and unit spacing in both directions, i.e., ∆x = ∆y = 1. Indexing cells by the integer c… view at source ↗
Figure 7
Figure 7. Figure 7: Extended results of view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative comparisons of rain field reconstructions on real CMLs links from OpenMRG dataset on three reconstruction tasks. The network of CMLs is depicted in red. 21 view at source ↗
Figure 9
Figure 9. Figure 9: Another set of qualitative comparisons of rain field reconstructions on real CMLs links from OpenMRG dataset on three reconstruction tasks. The network of CMLs is depicted in red. 22 view at source ↗

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

  1. [1]

    1” and “0

    doi: 10.1016/j.spasta.2017.12.001. Overeem, A., Leijnse, H., and Uijlenhoet, R. Country-wide rainfall maps from cellular communication networks.Pro- ceedings of the National Academy of Sciences, 110(8): 2741–2745, 2013. Overeem, A., Leijnse, H., and Uijlenhoet, R. Retrieval al- gorithm for rainfall mapping from microwave links in a cellular communication ...

    work page doi:10.1016/j.spasta.2017.12.001 2017

  2. [2]

    Indexing cells by the integer coordinates of their centers (c, r)∈Z 2 places the cell boundaries at c± 1 2 and r± 1 2; equivalently, cell centers lie at integer coordinates

    and unit spacing in both directions, i.e., ∆x= ∆y= 1 . Indexing cells by the integer coordinates of their centers (c, r)∈Z 2 places the cell boundaries at c± 1 2 and r± 1 2; equivalently, cell centers lie at integer coordinates. Therefore, the grid lines of the consideredH×Wdiscretization are located at x=c+ 1 2 , y=r+ 1 2 , c∈ {−1, . . . , W−1},andr∈ {−1...

    2006