Counterfactual Distribution Regression for Structured Inference

Arthur Gretton; Nicolo Colombo; Ricardo Silva; Soong M Kang

REVIEW 5 major objections 8 minor 21 references

Counterfactual Distribution Regression for Structured Inference

T0 review · 5 major / 8 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a distribution-to-distribution regression map, learned from past disruptions, can predict the effect of an unseen line closure on station exit-count distributions, without modeling the underlying causal mechanism.

desk verdict Genuinely new problem framing and a plausible application, but the consistency theorems rest on a wrong embedding rate and the evaluation hides the span-assumption risk behind selected high-severity disruptions. read the letter →

arxiv 1908.07193 v1 pith:SCL63M3R submitted 2019-08-20 cs.LG stat.APstat.MEstat.ML

Nicolo Colombo , Ricardo Silva , Soong M Kang , Arthur Gretton This is my paper · ORCID

classification cs.LGstat.APstat.MEstat.ML

keywords distributionregressionkernelmeanembeddingcounterfactualpredictionstructuredinterferencenetworkdisruptionsRKHSsimplex-constrainedsamplingmetropassengerflows

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 tries to establish a way to forecast how a complex system responds to a new external disruption—its running example is a metro network where a line closure changes passenger exit counts—by learning a single map from the counterfactual 'no-disruption' distribution to the disrupted distribution. The map is trained on natural-regime records plus a catalogue of past disruptions with features such as location and time, and then applied to an unseen disruption. If the map generalizes, planners could predict the whole distribution of passenger counts around affected stations from historical data alone, without building a detailed causal model of the system. The key move is to recast this distribution-to-distribution prediction as a regression between reproducing-kernel Hilbert space embeddings, with a simplex-constrained sampling step to turn the predicted embedding back into counts.

What carries the argument

The load-bearing object is the RKHS mean embedding $\mu_P = E_{X\sim P}[k(X,\cdot)]$, which represents a distribution by a point in a reproducing kernel Hilbert space. The argument rests on the identity $\mu_P = M_Q \alpha$: for every perturbation, the embedding of the disrupted distribution is assumed to be a fixed linear combination of the embeddings of $I$ hand-crafted input feature distributions derived from the natural regime and the disruption features, with a single coefficient vector $\alpha$ shared across all perturbations. This turns the infinite-dimensional map $\Psi$ into a finite linear regression that can be solved by least squares. The second mechanism is the simplex-constrained sampling scheme: given the predicted embedding, a convex program estimates mixture weights $\theta$ over a basis of station-specific marginal distributions, and samples are drawn from that mixture, with a consistency bound stated in the paper.

What would settle it

On a held-out set of disruptions, compute the residual norm $\|\hat{\mu}_{P_{\mathrm{perturbed}}} - \sum_i \hat{\alpha}_i \hat{\mu}_{Q_i}\|_H$ after fitting $\hat{\alpha}$ on other disruptions; if the residual is consistently large relative to the embedding norm of the observed distribution, the span assumption fails and predictions are forced toward the span. Equivalently, train the model on one half of the disruptions and test whether the same coefficients predict the other half no better than a baseline; if not, the fixed $\alpha$ assumption collapses.

Watch

Extended reading notes

Core claim

The central claim is that there exists a distribution-to-distribution regression map $\Psi$ such that, for a new perturbation with features $z^{(\mathrm{new})}$, the predicted distribution is $P_{\mathrm{perturbed}}^{(\mathrm{new})} = \Psi(P_{\mathrm{natural}}, z^{(\mathrm{new})}, A)$, where $P_{\mathrm{natural}}$ is the distribution of the system had no shock occurred. The paper estimates $\Psi$ by representing each distribution by its RKHS mean embedding and using a mixture-of-embeddings linear model: the embedding of the perturbed distribution is written as $\sum_{i=1}^I \alpha_i \hat{\mu}_{Q_i^{(\mathrm{new})}}$, with a fixed coefficient vector $\alpha$ learned from past disruptions. Because the output is an embedding rather than samples, the paper introduces a sampling step that fits a simplex-weighted mixture of station-specific marginal distributions to the predicted embedding and draws approximate samples from it. On out-of-sample disruptions in a metro network, the resulting predictive distributions are compared with a natural-regime baseline and with random mixtures.

Load-bearing premise

The assumption that any new disruption's effect is exactly a fixed linear combination of the hand-chosen summary features of the natural regime, with the same coefficients for every disruption; a novel disruption whose true effect falls outside these features cannot be corrected.

Editorial extensions

If this is right

Given natural-regime records and a library of past disruptions, the model can predict whole distributions of station exit counts for a closure at a previously unseen station.
No full causal graph or joint distribution over potential outcomes is required; the counterfactual distribution enters only as a covariate.
If the linear-span identity holds, coefficients learned from past disruptions transfer to new perturbations without retraining.
Predicted embeddings can be converted back into approximate samples through a single convex optimization, rather than an expensive per-sample optimization.
In the paper's out-of-sample evaluation, the proposed model often assigns higher likelihood to observed counts than the natural-regime baseline and random mixture models.

Reading between the lines

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

The same estimator should transfer to other graph-structured flow problems—road closures, power outages, epidemic spread on contact networks—where node-level counts under natural conditions and discrete shocks are observed.
The five hand-crafted features are a minimal dictionary; a natural extension is to learn a richer or data-adaptive dictionary so the linear-span assumption becomes an approximation with a controllable residual rather than an exact constraint.
Because the output is a full distribution rather than a point forecast, the method could be used to compute risk measures such as the probability that a station's exit count exceeds platform capacity, a quantity the paper does not explicitly evaluate.
A testable extension is to replace the fixed coefficient vector with one that depends on perturbation features, which would reveal which aspects of a disruption actually change how effects propagate.

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.

Desk Editor's Note

read the letter

Colombo et al. proposes something genuinely new: a distribution-to-distribution regression map that predicts the effect of a novel perturbation, with the perturbation described by node-level features on a network, and it demonstrates the idea on London Underground data. The formulation of the problem as a functional map between natural-regime input distributions and a perturbed output distribution is a real departure from standard distribution regression, and the simplex-constrained sampling from mean embeddings (Lemma 5/6) is a sensible alternative to herding. The out-of-sample evaluation over held-out disruptions, with baseline and random-model comparisons, is a reasonable design given the data constraints.

The soft spots are real. The theoretical rates are not established as written. Eq. (11) claims O(|D|^{-1}) for the mean-embedding error; the standard bound under a bounded kernel is O_p(|D|^{-1/2}). That error propagates into Lemmas 1-4, and the proof of Lemma 1 contains a bound after Eq. (61) that does not follow (the spectrum of M_Q^T M_Q appears on the wrong side of the inequality). So I would not rely on the consistency claims.

The bigger conceptual gap is the one the stress-test note flags: Eq. (23) assumes every perturbed embedding lies in the span of I hand-crafted feature embeddings with a single global alpha. The five Q_i in Eq. (45) are plausible summaries of the natural regime and disruption geometry, but there is no construction principle and no consistency under misspecification. A new disruption whose response has a component orthogonal to that span gets projected away, and the sampling step cannot recover it. This is a load-bearing modeling assumption, not a technical detail.

The empirical section also overstates what the data can show. The 20 disruptions are selected by the observable score in Eq. (49), which concentrates on high-severity cases where the linear model is most likely to look good. There is one observation per test disruption, the kernel bandwidth is tuned on the full dataset rather than per training fold, and no confidence intervals are reported. Code and data are not released. The empirical results are encouraging but not definitive.

Nothing about the citation pattern bothers me; the relevant distribution-regression and causal-inference literature is cited, and the Silva et al. (2015) connection is appropriate.

Who is this for? Researchers working on kernel mean embeddings and distribution regression, and people building decision-support tools for transport networks. It deserves a serious referee. I would send it to review, with the clear expectation of major revision: fix the rates or remove the consistency claims, justify or test the span assumption, and make the evaluation protocol more honest (report on all disruptions or justify the selection, use per-fold bandwidth selection, give uncertainty).

Referee Report

5 major / 8 minor

Summary. The paper introduces a distribution-to-distribution regression framework for predicting the effect of external perturbations on a structured system, using the motivating example of London Underground disruptions. The model assumes a meta-distribution over pairs of input feature distributions and output perturbed distributions, and estimates a linear operator on RKHS mean embeddings. A mixture-of-embeddings estimator (Eq. 22) is proposed, and a simplex-constrained optimization (Eq. 39) is used to sample from the predicted embedding. Consistency lemmas are stated for the estimators (Lemmas 1-4) and for the sampling procedure (Lemmas 5-6). The method is evaluated on a selection of 20 high-observable-score disruptions from the London Underground dataset, comparing log-likelihood and squared error against baseline and random models.

Significance. If the proposed method works as claimed, it would be a valuable tool for forecasting distributional effects of unseen disruptions in structured systems, and the strategy of reconstructing samples from a predicted mean embedding via simplex-constrained mixtures is an interesting contribution. The paper formulates a novel problem setting and provides a real-world application with Transport for London data. However, the theoretical guarantees are not established as stated: the embedding rate in Eq. (11) is incorrect, the proof of Lemma 1 contains a faulty inequality, and the main consistency results rely on a strong linear-span assumption (Eq. 23) for which no misspecification analysis is provided. The empirical evaluation is also narrow (20 selected, high-signal disruptions; one dataset; no comparison to alternative non-baseline regression models). The significance is therefore conditional on substantial technical revision and a more convincing empirical validation.

major comments (5)

[§2.3, Eq. (11)] The claim that the empirical mean embedding error satisfies ||µ_P - \hat µ_P||_H = O(1/|D_P|) is not correct for general characteristic kernels on unbounded or even bounded domains. For i.i.d. samples with a bounded kernel, the standard rate is O_p(1/√|D_P|) in RKHS norm (and O(1/√|D_P|) in expectation). This incorrect rate is used in the proofs of Lemmas 1-4 (e.g., Eqs. (54), (65), (75)), and therefore the consistency rates stated in Eqs. (15), (19), (24), and (31) are not established. The authors should either correct the rate to O_p(1/√D(K)) and propagate the consequences, or provide a proof under additional assumptions that justify a faster rate.
[Appendix, proof of Lemma 1, Eq. (62)] The inequality ∆(K) ≤ sqrt(σ_max(\hat N^T_extra \hat N_extra) σ_min(\hat M^T_Q \hat M_Q)) is not a valid operator norm bound. The term \hat N_extra (\hat M^T_Q \hat M_Q)^{-1} \hat M^T_Q should be bounded using the spectral norm of (\hat M^T_Q \hat M_Q)^{-1}, which involves 1/σ_min(\hat M^T_Q \hat M_Q), not σ_min. Furthermore, the bound in Eq. (63) overestimates the decay of σ_max(\hat N^T_extra \hat N_extra): for a K×K matrix, the spectral norm is at most the trace, which is O(K/D(K)), not O(1/D(K)). This proof needs to be rewritten carefully; as it stands, Lemma 1 is unsupported.
[§2.5, Lemma 3 and §3.2, Eq. (45)] The central assumption (23) states that every perturbed mean embedding lies exactly in the linear span of the five hand-crafted feature embeddings with a single coefficient vector α. No construction principle for the feature basis or consistency result under misspecification is provided. In the experiment, the five features (45) are heuristic summaries, and a novel disruption whose true effect has a component orthogonal to span{μ_{Q_i}^{(new)}} will be projected onto that span, with the error propagating to the sampling step and likelihood evaluation. The paper does not report the out-of-sample residual of the embedding fit or any measure of the span coverage. The selection of the 20 disruptions with the highest observable score (49), as shown in Figure 1, is likely to favor cases where the linear model is most accurate. This is a load-bearing gap: without a misspecification bound or residual check, the empirical claim of accurate prediction under unseen disruptions is not established.
[Appendix, proof of Lemma 4, Eq. (84)] The step ‖\bar w − \hat w^{(K)}‖ ≤ ‖\bar w − \hat α^{(K)}‖ does not follow from the convexity of the simplex constraint. The constrained least-squares solution \hat w is a projection of the unconstrained solution \hat α onto the feasible set only in the metric defined by the design matrix, not necessarily in the Euclidean norm or any norm used here. At minimum, the norm should be specified and the inequality proved. Without this, Lemma 4 is not supported.
[§3.3, Empirical results] The out-of-sample evaluation is compromised by the stated procedure for selecting the kernel bandwidth: 'the value of the kernel parameter ρ was obtained through cross-validation over the all sample and kept fixed over all training-testing instances.' This uses test data in the model-selection process, so the reported log-likelihoods and errors are not clean out-of-sample estimates. The authors should select ρ on each training fold only, or otherwise establish that the fixed ρ does not depend on test examples. This issue affects the credibility of the empirical claims.

minor comments (8)

[§2.3] Typo: 'Hibert space representation' should be 'Hilbert space representation'.
[§2.2, Eq. (5)] The smoothness condition is written as sup f(x) ≤ C‖f‖_H, but it should likely be |f(x)| ≤ C‖f‖_H; as written, it only gives an upper bound and does not guarantee the reproducing property holds in both directions.
[Definition 1, §2.4] The definition writes L_Ψ μ_P^{(k)} = μ_{ψ(Q^{(k)})}, but the underlying function Ψ is not explicitly defined in the definition; please clarify the notation and ensure ψ and Ψ are not confused.
[Eq. (17)] The formula for the one-parameter estimator is unclear: the trace notation and the placement of parentheses in (trace(\hat m_QQ)^{-1} (trace(\hat m_PQ)) are ambiguous. Please rewrite the expression.
[Lemma 6, Eq. (33)] The notation ε in the statement is introduced but not defined at the statement level; it is defined in the proof as the approximation error. Please make the definition explicit in the lemma.
[§3.2, after Eq. (45)] The index t and the summation limits in the definitions of X_i^{(k)} are not fully specified; for instance, E(Y_odjt) appears without an explicit definition of the expectation over the natural regime. Please add these details.
[§3.3] The text says 'ESPRC grant' in the acknowledgements; this should be 'EPSRC grant'.
[General] The paper contains numerous grammatical and typographical errors (e.g., 'reproducible kernel Hilbert spaces' appearing consistently as 'reproducible' rather than 'reproducing', and inconsistent use of 'reproducing' in other places). A thorough proofreading is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the prediction rule is a least-squares regression with out-of-sample evaluation; the exact-span condition is an explicit model assumption, not a derived output.

full rationale

The paper's central prediction step, Eq. (37), defines the predicted perturbed mean embedding as a linear combination of the new perturbation's feature embeddings using coefficients alpha_hat estimated by least squares in Eq. (22) from training disruptions. The alpha coefficients are fitted on a training fold and evaluated on disjoint held-out disruptions, so the out-of-sample predictions are not forced by the training fit. The exact-span assumption in Lemma 3, mu_P = M_Q alpha_bar, is stated explicitly as an assumption ('Assume that the meta-distribution ... can be represented by a unique RKHS linear operator L_QP such that mu_P = M_Q alpha_bar') and is used to prove a consistency rate for the least-squares estimator, not to define the target distribution. The hand-crafted feature distributions in Eq. (45) and the simplex-constrained sampling basis in Eq. (48) are modeling choices that restrict the output space and create a misspecification risk, but this is a statistical bias concern rather than a definitional or self-citational circularity. The self-citations in the paper are for background, motivation, and standard kernel-embedding results; none is used to justify the load-bearing assumptions or to forbid alternative models. The proof chain is therefore self-contained as a conditional consistency analysis plus an empirical evaluation, with no step in which a prediction reduces by construction to its own inputs.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The method rests on the assumption of an exact linear operator between embedding spaces (Eq 14) and on the hand-crafted feature family (Eq 36 and Eq 45). The consistency lemmas assume rather than derive the linear structure. The empirical rate O(1/n) for mean embeddings is stronger than the standard result, and the paper does not justify it. No new physical or conceptual entities are introduced beyond the perturbation map formulation.

free parameters (6)

Kernel bandwidth rho = not reported; cross-validated on full dataset
RKHS kernel parameter; selected by cross-validation over all data rather than per training fold (Section 3.3).
Number of input feature distributions I = 5
Hand-chosen in Section 3.2 with no sensitivity analysis.
Distance threshold xi = not specified
Defined in Eq (45) to separate feasible and infeasible paths; the value is not reported.
Scaling parameters lambda_r, R, c = not specified
Basis construction in Eq (48); R and c are not given numerical values.
Density smoothing bandwidth h = not specified
Used in Eq (51) for evaluation; the value is not reported.
Decay parameter beta (general method) = not specified
Eq (36) defines a feature family with beta > 0; not used in the experiment but part of the method.

assumptions (5)

ad hoc to paper There exists a unique RKHS linear operator L_true such that mu_P = L_true mu_Q for all (P,Q) from the meta-distribution (Eq 14).
Central modeling assumption; if false, the estimated operator has no interpretation and prediction is unconstrained by theory.
domain assumption The 35 days of natural-regime records are exchangeable samples from the counterfactual no-disruption distribution for any disruption day.
Used to build P_natural; day-to-day variation and trends would violate this.
domain assumption The graph G and adjacency matrix A are given and fixed; the graph encodes physical adjacency, not causal structure.
Required for the spatial features in Eq (36) and Eq (45).
standard math The kernel k is characteristic, so the mean embedding is injective (Eq 9).
Standard RKHS result used to justify distribution embeddings.
ad hoc to paper Empirical mean embedding error is O(1/|D|) (Eq 11).
The paper assumes a rate O(1/n) for the RKHS embedding error, which is stronger than the standard O_p(1/sqrt(n)); this underpins Lemmas 1 to 4.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Counterfactual Distribution Regression for Structured Inference." pith.science (2026). https://pith.science/paper/SCL63M3R

@misc{pith2026190807193,
  author       = {Pith},
  title        = {Pith review of: Counterfactual Distribution Regression for Structured Inference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SCL63M3R}},
  note         = {Machine review of arXiv:1908.07193}
}

read the original abstract

We consider problems in which a system receives external \emph{perturbations} from time to time. For instance, the system can be a train network in which particular lines are repeatedly disrupted without warning, having an effect on passenger behavior. The goal is to predict changes in the behavior of the system at particular points of interest, such as passenger traffic around stations at the affected rails. We assume that the data available provides records of the system functioning at its "natural regime" (e.g., the train network without disruptions) and data on cases where perturbations took place. The inference problem is how information concerning perturbations, with particular covariates such as location and time, can be generalized to predict the effect of novel perturbations. We approach this problem from the point of view of a mapping from the counterfactual distribution of the system behavior without disruptions to the distribution of the disrupted system. A variant on \emph{distribution regression} is developed for this setup.

Figures

Figures reproduced from arXiv: 1908.07193 by the authors.

**Figure 2.** Logarithm of the negative log-likelihood of the disruption-day exit counts, () [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗

**Figure 3.** Logarithm of the square error kEY ∼P (k) model (Y ) − yˆk 2kyˆk −2 , with yˆ ∈ DP (k) perturbed and P (k) model being the baseline model (left, x-axis), a random model (right, x-axis) and the proposed model (left and right, y-axis). Red markers highlight cases where the proposed model outperforms the models to which it is compared [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Exit counts probability distributions (selection 1): disruptions and stations where [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]

**Figure 5.** Figure 5: Exit counts probability distributions (selection 2): disruptions stations where the [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 18 canonical work pages

[1]

Estimating spatial effects

Peter M Aronow and Cyrus Samii. Estimating spatial effects. Unpublished manuscript, 2017

work page 2017
[2]

Estimating average causal effects under general interference, with application to a social network experiment

Peter M Aronow, Cyrus Samii, et al. Estimating average causal effects under general interference, with application to a social network experiment. The Annals of Applied Statistics, 11 0 (4): 0 1912--1947, 2017

work page 1912
[3]

Cooper and C

G. Cooper and C. Yoo. Causal discovery from a mixture of experimental and observational data. Proceedings of the 15th conference on Uncertainty in Artificial Intelligencem (UAI-1999), pages 116--125, 1999

work page 1999
[4]

T. Dunning. Natural Experiments in the Social Sciences: A Design-Based Approach. Cambridge University Press, 2012

work page 2012
[5]

Dimensionality reduction for supervised learning with reproducing kernel hilbert spaces

Kenji Fukumizu, Francis R Bach, and Michael I Jordan. Dimensionality reduction for supervised learning with reproducing kernel hilbert spaces. Journal of Machine Learning Research, 5 0 (Jan): 0 73--99, 2004

2004
[6]

Imbens and D

G. Imbens and D. Rubin. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction. Cambridge University Press, 2015

work page 2015
[7]

Predicting the future behavior of a time-varying probability distribution

Christoph H Lampert. Predicting the future behavior of a time-varying probability distribution. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 942--950, 2015

work page 2015
[8]

Bayesian approaches to distribution regression

Ho Chung Leon Law, D Sutherland, Dino Sejdinovic, and Seth Flaxman. Bayesian approaches to distribution regression. population, 2: 0 3, 2018

work page 2018

Show all 21 references

[9]

Kernel mean embedding of distributions: A review and beyond

Krikamol Muandet, Kenji Fukumizu, Bharath Sriperumbudur, Bernhard Sch \"o lkopf, et al. Kernel mean embedding of distributions: A review and beyond. Foundations and Trends in Machine Learning , 10 0 (1-2): 0 1--141, 2017

2017
[10]

Novershtern, A

N. Novershtern, A. Regev, and N. Friedman. Physical module networks: an integrative approach for reconstructing transcription regulation. Bioinformatics, 27: 0 177--185, 2011

2011
[11]

J. Pearl. Causality: M odels, R easoning and I nference . Cambridge University Press, 2000

2000
[12]

Peters, P

J. Peters, P. Buhlmann, and N. Meinshausen. Causal inference using invariant prediction: identification and confidence intervals (with discussion). Journal of the Royal Statistical Society: Series B, 78: 0 947--1012, 2016

2016
[13]

Silva, S

R. Silva, S. M. Kang, and E. M. Airoldi. Predicting traffic volumes and estimating the effects of shocks in massive transportation systems. Proceedings of the National Academy of Sciences, 112: 0 5643--5648, 2015

2015
[14]

A hilbert space embedding for distributions

Alex Smola, Arthur Gretton, Le Song, and Bernhard Sch \"o lkopf. A hilbert space embedding for distributions. In International Conference on Algorithmic Learning Theory, pages 13--31. Springer, 2007

2007
[15]

S. E. F. Spencer, S. M. Hill, and S. Mukherjee. Inferring network structure from interventional time-course experiments. Annals of Applied Statistics, 9: 0 507--524, 2015

2015
[16]

Spirtes, C

P. Spirtes, C. Glymour, and R. Scheines. Causation, P rediction and S earch . Cambridge University Press, 2000

2000
[17]

Hilbert space embeddings and metrics on probability measures

Bharath K Sriperumbudur, Arthur Gretton, Kenji Fukumizu, Bernhard Sch \"o lkopf, and Gert RG Lanckriet. Hilbert space embeddings and metrics on probability measures. Journal of Machine Learning Research, 11 0 (Apr): 0 1517--1561, 2010

2010
[18]

On the empirical estimation of integral probability metrics

Bharath K Sriperumbudur, Kenji Fukumizu, Arthur Gretton, Bernhard Sch \"o lkopf, Gert RG Lanckriet, et al. On the empirical estimation of integral probability metrics. Electronic Journal of Statistics, 6: 0 1550--1599, 2012

2012
[19]

Kernels on sample sets via nonparametric divergence estimates

Dougal J Sutherland, Liang Xiong, Barnab \'a s P \'o czos, and Jeff Schneider. Kernels on sample sets via nonparametric divergence estimates. arXiv preprint arXiv:1202.0302, 2012

2012 arXiv
[20]

Learning theory for distribution regression

Zolt \'a n Szab \'o , Bharath Sriperumbudur, Barnab \'a s P \'o czos, and Arthur Gretton. Learning theory for distribution regression. Journal of Machine Learning Research, 17 0 (152): 0 1--40, 2016

2016
[21]

Causal discovery from changes

Jin Tian and Judea Pearl. Causal discovery from changes. In Proceedings of the Seventeenth conference on Uncertainty in artificial intelligence, pages 512--521. Morgan Kaufmann Publishers Inc., 2001

2001

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Estimating spatial effects

Peter M Aronow and Cyrus Samii. Estimating spatial effects. Unpublished manuscript, 2017

work page 2017

[2] [2]

Estimating average causal effects under general interference, with application to a social network experiment

Peter M Aronow, Cyrus Samii, et al. Estimating average causal effects under general interference, with application to a social network experiment. The Annals of Applied Statistics, 11 0 (4): 0 1912--1947, 2017

work page 1912

[3] [3]

Cooper and C

G. Cooper and C. Yoo. Causal discovery from a mixture of experimental and observational data. Proceedings of the 15th conference on Uncertainty in Artificial Intelligencem (UAI-1999), pages 116--125, 1999

work page 1999

[4] [4]

T. Dunning. Natural Experiments in the Social Sciences: A Design-Based Approach. Cambridge University Press, 2012

work page 2012

[5] [5]

Dimensionality reduction for supervised learning with reproducing kernel hilbert spaces

Kenji Fukumizu, Francis R Bach, and Michael I Jordan. Dimensionality reduction for supervised learning with reproducing kernel hilbert spaces. Journal of Machine Learning Research, 5 0 (Jan): 0 73--99, 2004

2004

[6] [6]

Imbens and D

G. Imbens and D. Rubin. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction. Cambridge University Press, 2015

work page 2015

[7] [7]

Predicting the future behavior of a time-varying probability distribution

Christoph H Lampert. Predicting the future behavior of a time-varying probability distribution. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 942--950, 2015

work page 2015

[8] [8]

Bayesian approaches to distribution regression

Ho Chung Leon Law, D Sutherland, Dino Sejdinovic, and Seth Flaxman. Bayesian approaches to distribution regression. population, 2: 0 3, 2018

work page 2018