Deep Causal Inference for Point-referenced Spatial Data with Continuous Treatments

David Carlson; Lei Duan; Yiling Liu; Zach Calhoun; Ziyang Jiang

REVIEW 5 major objections 6 minor 12 references

Deep Causal Inference for Point-referenced Spatial Data with Continuous Treatments

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Neural-network spatial regression beats linear models at estimating direct, indirect, and total causal effects with continuous treatments and local spillover.

desk verdict Useful integration of NN outcome regression, approximate GP, and GPS weights for spatial causal inference, but the headline claim of significant improvement is not backed by the reported experiments. read the letter →

arxiv 2412.04285 v1 pith:W2LXWH6P submitted 2024-12-05 cs.LG

Ziyang Jiang , Zach Calhoun , Yiling Liu , Lei Duan , David Carlson This is my paper · ORCID

classification cs.LG MSC 62D2062M3068T07

keywords spatialcausalinferencecontinuoustreatmentsinterferenceunobservedconfoundingneuralnetworksGaussianprocessgeneralizedpropensityscorepoint-referenceddata

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

The paper sets out to replace linear spatial regression with a neural-network version for causal inference on point-referenced data with continuous treatments, where a unit's outcome can be affected by treatments at neighboring locations (spatial interference) and by hidden spatial confounders. The proposed model keeps each treatment's direct effect linear, learns the indirect spillover effects and the effects of measured confounders with neural networks, models the unobserved spatial confounder as a sample from an approximate Gaussian process, and uses generalized-propensity-score balancing weights to handle the fact that each outcome is observed under only one treatment value. The paper argues this matters because environmental data derived from satellite imagery often have nonlinear, non-stationary relationships that linear spatial models cannot capture. Its experiments on synthetic, semi-synthetic, and real-world heat-island data report that the neural-network models estimate direct, indirect, and total effects with lower error than the linear spatial regression baseline, and that the network-based indirect-effect maps agree better with expected land-cover behavior.

What carries the argument

The load-bearing object is the additive decomposition in Eq. (9). It separates the outcome into a linear direct effect $\alpha_m T_{i,m}$ per treatment, neural-network functions $f_{\theta_{t_m}}$ for indirect effects of neighboring treatments and $g_{\theta_x}$ for measured confounders, an approximate-GP term $U_\phi(s_i)$ for the unobserved spatial confounder, and noise. The approximate GP is built through the Implicit Composite Kernel: a low-rank inducing-point approximation of an isotropic covariance matrix, followed by a Cholesky decomposition, so that the kernel parameters and the network weights can be optimized jointly by stochastic gradient descent. To estimate effects when outcomes are partially observed, the paper multiplies predicted outcomes by generalized-propensity-score balancing weights $w_m^{(i)} = f_{T_m}(t_m^{(i)}) / f_{T_m \mid X,s}(t_m^{(i)} \mid X^{(i)}, s^{(i)})$, with self-normalization to guard against propensity overfitting.

What would settle it

Simulate a semi-synthetic dataset whose true spillover radius is larger than the model's fixed neighborhood $S$ and whose unobserved confounder includes short-range variation not captured by the chosen kernel lengthscale; if the estimation error of the direct, indirect, and total effects grows with the size of this mismatch, the identifying assumptions have been falsified.

Watch

Extended reading notes

Core claim

The central claim is that the model in Eq. (9) — $Y_i = \sum_{m=1}^M \alpha_m T_{i,m} + \sum_{m=1}^M f_{\theta_{t_m}}(\bar{T}_{i,m}) + g_{\theta_x}(X_i) + U_\phi(s_i) + \varepsilon$ — estimates causal effects more accurately than the linear spatial regression model of Eq. (2) when interference and confounding are nonlinear. The direct effect of each treatment assigned to a unit is kept as a linear term so that it is not overshadowed by the many neighboring treatments; the indirect effect from treatments inside a local neighborhood $S$ and the effect of measured confounders are learned by neural networks; and $U_\phi(s_i)$ is a differentiable approximate-GP term representing unobserved spatial confounding. On synthetic and semi-synthetic data the paper reports that this specification, especially with the GP term and balancing weights, reduces estimation error for direct, indirect, and total effects compared with the linear baseline. In the real-world study, the CNN- and U-Net-based models produce indirect effects of NDVI on urban temperature whose magnitude decreases with building intensity, matching the paper's physical expectation, while the linear model shows the pattern less clearly.

Load-bearing premise

The whole estimator rests on local interference — that treatments outside the chosen neighborhood $S$ have negligible effect — and on latent ignorability given observed confounders plus a smoothly varying spatial unobserved confounder; if spillover reaches beyond $S$ or the hidden confounder changes sharply within a neighborhood, the estimates are biased.

Editorial extensions

If this is right

Spatial causal inference can move beyond linear models: high-dimensional inputs such as satellite imagery can feed the indirect-effect and confounder networks while the direct effect stays interpretable.
Including the approximate spatial-GP term $U_\phi(s_i)$ reduces bias in direct, indirect, and total effect estimates when unmeasured spatial confounding is present, as shown in the paper's synthetic and semi-synthetic experiments.
Balancing weights from the generalized propensity score, with self-normalization, make the effect estimates usable when each unit is observed under a single continuous treatment value.
The CNN and U-Net variants give more spatially detailed indirect-effect maps, so decisions about vegetation and albedo interventions can be informed by where the spillover cooling is strongest.

Reading between the lines

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

If the central claim holds, the same architecture pattern — a linear target effect, network nuisance terms, and a spatial-process unobserved confounder — could transfer to other interference structures such as river networks or commuting zones, where 'neighboring' is defined by connectivity rather than Euclidean distance.
A natural stress test the paper does not run is to vary the neighborhood window $S$ and the kernel lengthscale; estimates that shift markedly when either changes would reveal that the local-interference and smooth-confounder assumptions, not just the network fitting, are doing the work.
Extending the balancing-weight step to include neighboring treatments in the propensity model would be a testable modification, since the paper currently assumes treatments are spatially independent.

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

Colleague,

The paper earns a serious look: it assembles an existing toolbox—CNN/U-Net for spatial treatments, approximate GP for unobserved spatial confounding, GPS balancing weights—into a single framework for point-referenced data with continuous treatments, and it ships code. That integration is genuinely new and the real-world Durham case study (NDVI/albedo → temperature) is a good stress test for the method.

But the headline claim, that NN models significantly outperform linear models, is not actually supported by the evidence as reported. The semi-synthetic Table 2 has a single run with no error bars or significance test; that cannot carry a "significantly" claim. The synthetic Table 1 shows overlapping standard deviations for DE and TE between the linear-with-Uphi and NN-with-Uphi models, so even there the advantage is not clear-cut. And the GPS in Section 3.3 is said to be estimated with a logistic regression model—that is a binary classifier, not a density estimator for a continuous treatment. Unless the treatment was discretized (not stated), the balancing weights in Eq. (10) are not valid as written. This is the kind of thing a referee should force the authors to clarify or fix.

There is also a mild circularity in the semi-synthetic evaluation: ground truth is generated with a nonlinear spline plus a random NN, exactly the functional class the NN outcome model is designed to fit. The linear baseline is then at a structural disadvantage. That doesn't sink the method, but it means the comparison is more favorable than a neutral benchmark would be. The local interference and isotropic GP assumptions are acknowledged and are standard for this literature; they're limitations, not flaws.

The paper is worth refereeing. The problem is real, the framework is useful, and the issues are fixable: rerun with multiple seeds and report error bars, correct or clarify the GPS implementation, benchmark against the nearest nonlinear spatial methods (Pollmann, Tec, Giffin). I would not cite it in its current form, but I'd happily read a revised version. Send it to peer review rather than desk-reject.

Referee Report

5 major / 6 minor

Summary. The manuscript proposes a neural-network-based spatial regression model for point-referenced causal inference with continuous treatments. The model in Eq. (9) keeps direct effects linear, represents indirect effects f_θt and measured-confounder effects g_θx with deep networks, and models the unobserved spatial confounder U_φ(s) via a low-rank approximate Gaussian process from Jiang et al. (2024). Causal effects are defined under a local-interference assumption, and balancing weights from a generalized propensity score are used to handle partially observed outcomes. The paper reports synthetic, semi-synthetic, and real-world satellite-image experiments and claims in the abstract that NN-based models significantly outperform linear spatial regression models in estimating causal effects.

Significance. If the central claim were fully supported, the paper would provide a usable nonlinear alternative to linear spatial causal inference on point-referenced data, extending the scope to continuous treatments and high-dimensional inputs. The paper has concrete strengths: code is released; the approximate-GP component is imported from a separately published and validated framework; the synthetic experiment is repeated five times with means and standard deviations; and the real-world case study uses publicly available satellite and temperature data. However, the evidence for the headline claim is currently incomplete. Table 2 reports only point estimates with no error bars or replication information, and the GPS estimation step described in Section 3.3 appears misspecified for continuous treatments. The identification assumptions are standard for this literature, but their sensitivity is not explored. These issues are local and fixable within the manuscript's scope.

major comments (5)

[§3.3, Eq. (10)] The manuscript states that the generalized propensity score (GPS) for a continuous treatment is estimated 'using a logistic regression model.' Logistic regression is a model for a binary probability, not for a continuous conditional density, and the paper does not state that the treatment is discretized. Since Eq. (10) requires the conditional density f_{T_m|X,s} in the denominator, the balancing weights are not well defined as written, and the DE/IE/TE estimates in Tables 1–2 and the real-world analysis inherit this issue. Please either replace the GPS estimator with an appropriate continuous-density model (e.g., a flexible regression with a kernel density estimate of residuals or a heteroscedastic normal model), or explicitly describe a discretization scheme and assess the resulting coarsening bias.
[Abstract; Tables 1 and 2] The abstract claims that NN-based models 'significantly outperform' linear models, but the reported statistics do not support the word 'significantly.' In Table 1, the NN model with Uφ and the linear model with Uφ have overlapping standard deviations for DE (0.096±0.016 vs 0.108±0.012) and TE (0.095±0.016 vs 0.115±0.018), and the paper reports no significance tests. Table 2 reports a single set of point estimates with no number of repeats, no error bars, and no significance test. Please add repeated runs with standard errors or confidence intervals and explicit tests, or restrict the claim to 'lower estimation error in these experiments.'
[§6.1, Eq. (12)] The semi-synthetic ground truth is generated with f_T as a piecewise cubic spline and f_X as a randomly initialized neural network, which are the same functional classes as the NN components f_θt and g_θx in Eq. (9). Consequently, Table 2 measures performance inside the model class the NN is designed to fit and should be framed as a favorable-case evaluation; it does not by itself establish that the NN estimator is more accurate for arbitrary nonlinear spatial mechanisms. Adding a second semi-synthetic scenario outside the NN function class, for example a discontinuous or nonstationary mechanism, would substantially strengthen the comparative claim.
[§2, Assumption 2; §3.2, Table 5] The identification of direct, indirect, and total effects rests on local interference with a fixed neighborhood S and on latent ignorability given an unobserved confounder drawn from an isotropic GP with a specific kernel. Neither assumption is probed: the neighborhood size d_S is set to 51 in the semi-synthetic study without a sensitivity analysis, and the RBF lengthscale in Table 5 is fixed at 0.005 for semi-synthetic data and 100 for real data. If spillover extends beyond S, or if the unobserved confounder varies at a scale shorter than the kernel lengthscale, the estimated effects are biased. Please add a sensitivity analysis over S and kernel lengthscale, or state explicitly that these are maintained identifying assumptions rather than quantities validated by the experiments.
[§6.2, Table 3] The real-world evaluation reports R² and MAE for temperature prediction, which are measures of outcome prediction accuracy and not of causal effect accuracy. The causal-effect claim in this section rests on the qualitative comparison in Figure 3, which is suggestive but is not a quantitative validation. Please state this distinction explicitly and avoid presenting the prediction statistics in Table 3 as evidence about the accuracy of the estimated causal effects.

minor comments (6)

[§5, last paragraph] The text says 'the estimation error on DE, TE, and TE' and should read 'DE, IE, and TE.'
[§7] The first sentence of the Limitations section reads 'One limitation of the proposed lies in the model form of (9)' and is missing the word 'method' or 'framework.'
[§5, Eq. (11)] The notation d_i is introduced as a row of the N×N matrix D, but T_i contains only the treatments of two neighboring nodes; please specify how d_i is restricted to the neighborhood or redefine the notation to avoid ambiguity.
[§3.2, Table 5] The kernel hyperparameters σ, l, and ε are listed as fixed values in Table 5, while Section 3.2 says φ includes the kernel parameters and is learned during training; please clarify which parameters are estimated and which are prespecified.
[§2, Assumption 3] For continuous treatments, the positivity assumption should be stated in terms of a conditional density that is bounded away from zero on the support of the treatment; the current statement p(T_m=t_m|X,U)>0 for all t_m ∈ R is stronger than needed and may be impossible for bounded treatments such as NDVI.
[§3.3] The description of self-normalization is ambiguous: the text says each predicted outcome is weighed by 'a self-normalization term 1/N Σ_i w_i^(m)', but the standard self-normalized importance-sampling estimator divides by the average weight rather than multiplying by it; please clarify the exact estimator implemented in the code.

Circularity Check

0 steps flagged · score 1.0 of 10

No derivation reduces to its inputs; the only self-citation (the ICK framework) is separate published support, so any circularity is minor rather than load-bearing.

full rationale

We examined Eqs. (2), (9), (3)-(8), and the semi-synthetic generation in Eqs. (11)-(12). The causal-effect estimates are post-processing of the fitted outcome model; no target effect is used as a fitted input, and no fitted parameter is renamed as a prediction. The direct-effect term alpha_m T_i,m in Eq. (9) is a modeling restriction, not a definition of the direct effect, which is independently defined in Eqs. (3) and (6). The approximate-GP component U_phi is imported from Jiang et al. (2024), a self-citation, but that work is a separate TMLR publication with stated theoretical and empirical support; the central NN-versus-linear claim does not reduce to the validity of that citation. The semi-synthetic benchmark generates outcomes using nonlinear functions and a spatial unobserved confounder, which is favorable to the proposed NN-plus-GP model, but the ground truth is not the fitted function, so the comparison is a favorable evaluation design rather than a by-construction equivalence. Concerns about using logistic regression for a continuous generalized propensity score and the absence of error bars in Table 2 are correctness and statistical-evidence issues, not circularity. No specific equation reduces to its own input by construction, and no loaded self-citation chain forces the reported result; therefore the circularity score is appropriately low.

Assumptions & free parameters 5 free parameters · 8 assumptions · 0 invented entities

The empirical comparison rests on identification assumptions and data-generation choices more than on a formal derivation. The model's trained weights are not listed as free parameters; the hand-chosen quantities are the kernel hyperparameters, the unspecified inducing-point count, and the unspecified kernel density bandwidth, all of which affect the estimated causal effects. No new entities are introduced.

free parameters (5)

RBF kernel lengthscale for U_phi = l = 0.5 (synthetic), 0.005 (semi-synthetic), 100 (real-world)
Chosen by hand per dataset in Table 5; controls the spatial smoothness of the unobserved confounder and is not learned or cross-validated in a reported way.
RBF kernel noise level for U_phi = epsilon = 0.5 (synthetic), 0.1 (semi-synthetic and real-world)
Chosen by hand in Table 5.
RBF kernel variance for U_phi = sigma = 1 (all datasets)
Chosen by hand in Table 5; fixed to 1.
Number of inducing points q for the low-rank GP approximation = not stated
Section 3.2 defines K approx K_nq K_q^-1 K_nq^T and requires q, but the paper never reports q or its selection; this is a hidden degree of freedom in the approximate GP.
KDE bandwidth for the marginal treatment density = not stated
Section 3.3 says the marginal density in the balancing weight is estimated by kernel smoothing, but the bandwidth is not specified, which affects the weights.

assumptions (8)

domain assumption Consistency: the observed outcome is the potential outcome for the observed treatments in S (Assumption 1).
Standard potential-outcomes assumption invoked in Section 2.1; needed to equate observed Y with the potential outcome under the observed treatment.
domain assumption Latent ignorability: Y is independent of T_m given X and U for all m (Assumption 2).
The central identification assumption; it requires that after conditioning on observed confounders and the purely spatial unobserved confounder U there is no residual confounding. Not tested in the paper.
domain assumption Positivity: p(T_m = t_m | X, U) > 0 for all m and t_m (Assumption 3).
Required for generalized propensity score weighting; unverifiable for continuous treatments in finite samples.
domain assumption Local interference: outcomes depend on treatments only inside a pre-specified neighboring area S; treatments outside S have negligible effect.
Introduced before Eq. (2) and in Section 2; if this is wrong, indirect and total effect estimates are biased. S is chosen by hand.
domain assumption The unobserved confounder U is a purely spatial term, constant over each spatial region and sampled from a zero-mean Gaussian process with isotropic covariance.
Section 2 and Section 3.2; if U varies within a neighborhood or has non-isotropic structure, the GP model cannot capture it.
domain assumption The low-rank implicit composite kernel approximation from Jiang et al. (2024) faithfully approximates a full GP.
Section 3.2 imports this approximation as established; the paper provides no new proof and relies on the cited TMLR paper.
domain assumption The GPS balancing weights correctly adjust for partial outcome observation under the estimated GPS model.
Section 3.3; the estimator in Eq. (10) is asserted following Robins et al. and Assaad et al., but the paper does not verify the GPS specification, especially the unusual logistic-regression fit for a continuous treatment.
domain assumption Treatments are assumed not to be spatially related to each other.
Section 3.3 states "we do not assume any spatial relationships among treatment variables"; if treatments are spatially autocorrelated, the balancing weight formula may not correct the resulting confounding.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Deep Causal Inference for Point-referenced Spatial Data with Continuous Treatments." pith.science (2026). https://pith.science/paper/W2LXWH6P

@misc{pith2026241204285,
  author       = {Pith},
  title        = {Pith review of: Deep Causal Inference for Point-referenced Spatial Data with Continuous Treatments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W2LXWH6P}},
  note         = {Machine review of arXiv:2412.04285}
}

read the original abstract

Causal reasoning is often challenging with spatial data, particularly when handling high-dimensional inputs. To address this, we propose a neural network (NN) based framework integrated with an approximate Gaussian process to manage spatial interference and unobserved confounding. Additionally, we adopt a generalized propensity-score-based approach to address partially observed outcomes when estimating causal effects with continuous treatments. We evaluate our framework using synthetic, semi-synthetic, and real-world data inferred from satellite imagery. Our results demonstrate that NN-based models significantly outperform linear spatial regression models in estimating causal effects. Furthermore, in real-world case studies, NN-based models offer more reasonable predictions of causal effects, facilitating decision-making in relevant applications.

Figures

Figures reproduced from arXiv: 2412.04285 by the authors.

**Figure 2.** Geospatial datasets used for semi-synthetic and real-world experiments. Land cover classes are provided in [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

**Figure 3.** Distribution of the indirect effect on temperature from NDVI for both linear and NN-based models, where [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Visualization of the direct and indirect effects from NDVI and albedo on the surface air temperature in the [PITH_FULL_IMAGE:figures/full_fig_p015_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 7 canonical work pages

[8]

Instrumental variables, spatial confounding and interference

Andrew Giffin, Brian J Reich, Shu Yang, and Ana G Rappold. Instrumental variables, spatial confounding and interference. arXiv preprint arXiv:2103.00304,

arXiv
[10]

Estimating causal effects using a multi-task deep ensemble

12 arXiv Template A PREPRINT Ziyang Jiang, Zhuoran Hou, Yiling Liu, Yiman Ren, Keyu Li, and David Carlson. Estimating causal effects using a multi-task deep ensemble. In International Conference on Machine Learning, pages 15023–15040. PMLR, 2023b. Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980,

arXiv
[12]

The Sentinel-2 data is collected using Google Earth using the least cloudy data within 2 months of the date of the temperature collection

based on Sentinel-2 surface reflectance data. The Sentinel-2 data is collected using Google Earth using the least cloudy data within 2 months of the date of the temperature collection. 14 arXiv Template A PREPRINT Direct effect from NDVI Direct effect from albedo Indirect effect from NDVI Indirect effect from albedo Linear modelfθt: CNN, gθx: MLP fθt: U-N...

work page 2021
[2000]

Counterfactual representation learning with balancing weights

Serge Assaad, Shuxi Zeng, Chenyang Tao, Shounak Datta, Nikhil Mehta, Ricardo Henao, Fan Li, and Lawrence Carin. Counterfactual representation learning with balancing weights. In International Conference on Artificial Intelligence and Statistics, pages 1972–1980. PMLR,

work page 1972
[2004]

Twitter sentiment analysis of covid vaccines

Wenbo Zhu and Tiechuan Hu. Twitter sentiment analysis of covid vaccines. In 2021 5th International Conference on Artificial Intelligence and Virtual Reality (AIVR), pages 118–122,

work page 2021
[2015]

Causal inference for spatial treatments

Michael Pollmann. Causal inference for spatial treatments. arXiv preprint arXiv:2011.00373,

arXiv 2011
[2018]

Optimizing distributed networking with big data scheduling and cloud computing

Wenbo Zhu. Optimizing distributed networking with big data scheduling and cloud computing. In International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022), volume 12303, pages 23–28. SPIE,

work page 2022
[2019]

Spatial causal inference in the presence of unmeasured confounding and interference

Georgia Papadogeorgou and Srijata Samanta. Spatial causal inference in the presence of unmeasured confounding and interference. arXiv preprint arXiv:2303.08218,

arXiv

Show all 12 references

[2020]

13 arXiv Template A PREPRINT Table 4: Land cover classes corresponding to the color codes in Figure 2c. Code Land Cover Type 11 Open Water 21 Developed, Open Space 22 Developed, Low Intensity 23 Developed, Medium Intensity 24 Developed, High Intensity 31 Barren Land 41 Deciduo...

2024
[2021]

U-net: Convolutional networks for biomedical image segmenta- tion

Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmenta- tion. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, par...

2015
[2022]

Artificial intelligence aspect of transportation analysis using large scale systems

Tiechuan Hu, Wenbo Zhu, and Yuqi Yan. Artificial intelligence aspect of transportation analysis using large scale systems. In Proceedings of the 2023 6th Artificial Intelligence and Cloud Computing Conference, pages 54–59,

2023
[2023]

Causal mediation analysis with multi-dimensional and indirectly observed mediators

Ziyang Jiang, Yiling Liu, Michael H Klein, Ahmed Aloui, Yiman Ren, Keyu Li, Vahid Tarokh, and David Carl- son. Causal mediation analysis with multi-dimensional and indirectly observed mediators. arXiv preprint arXiv:2306.07918, 2023a. Paul R Rosenbaum and Donald B Rubin. The c...

arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [8]

Instrumental variables, spatial confounding and interference

Andrew Giffin, Brian J Reich, Shu Yang, and Ana G Rappold. Instrumental variables, spatial confounding and interference. arXiv preprint arXiv:2103.00304,

arXiv

[2] [10]

Estimating causal effects using a multi-task deep ensemble

12 arXiv Template A PREPRINT Ziyang Jiang, Zhuoran Hou, Yiling Liu, Yiman Ren, Keyu Li, and David Carlson. Estimating causal effects using a multi-task deep ensemble. In International Conference on Machine Learning, pages 15023–15040. PMLR, 2023b. Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980,

arXiv

[3] [12]

The Sentinel-2 data is collected using Google Earth using the least cloudy data within 2 months of the date of the temperature collection

based on Sentinel-2 surface reflectance data. The Sentinel-2 data is collected using Google Earth using the least cloudy data within 2 months of the date of the temperature collection. 14 arXiv Template A PREPRINT Direct effect from NDVI Direct effect from albedo Indirect effect from NDVI Indirect effect from albedo Linear modelfθt: CNN, gθx: MLP fθt: U-N...

work page 2021

[4] [2000]

Counterfactual representation learning with balancing weights

Serge Assaad, Shuxi Zeng, Chenyang Tao, Shounak Datta, Nikhil Mehta, Ricardo Henao, Fan Li, and Lawrence Carin. Counterfactual representation learning with balancing weights. In International Conference on Artificial Intelligence and Statistics, pages 1972–1980. PMLR,

work page 1972

[5] [2004]

Twitter sentiment analysis of covid vaccines

Wenbo Zhu and Tiechuan Hu. Twitter sentiment analysis of covid vaccines. In 2021 5th International Conference on Artificial Intelligence and Virtual Reality (AIVR), pages 118–122,

work page 2021

[6] [2015]

Causal inference for spatial treatments

Michael Pollmann. Causal inference for spatial treatments. arXiv preprint arXiv:2011.00373,

arXiv 2011

[7] [2018]

Optimizing distributed networking with big data scheduling and cloud computing

Wenbo Zhu. Optimizing distributed networking with big data scheduling and cloud computing. In International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022), volume 12303, pages 23–28. SPIE,

work page 2022

[8] [2019]

Spatial causal inference in the presence of unmeasured confounding and interference

Georgia Papadogeorgou and Srijata Samanta. Spatial causal inference in the presence of unmeasured confounding and interference. arXiv preprint arXiv:2303.08218,

arXiv