Pith. sign in

REVIEW 4 major objections 5 minor 50 references

Dimension-independent rates for structured neural density estimation

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read For densities Markov to a graph, a neural estimator reaches rate $n^{-1/(4+r)}$ and the optimal $L^1$ rate is $n^{-1/(2+r)}$, so clique size is effective dimension.

desk verdict Novel effective-dimension insight for MRF density estimation, but the main theorem as stated is an oracle statement because the network class is chosen using the unknown density. read the letter →

arxiv 2411.15095 v1 pith:D47M4HZI submitted 2024-11-22 stat.ML cs.CVcs.LGmath.STstat.TH

Robert A. Vandermeulen , Wai Ming Tai , Bryon Aragam This is my paper · ORCID
classification stat.MLcs.CVcs.LGmath.STstat.TH MSC 62G0762G2068T07
keywords densityestimationMarkovrandomfielddimension-independentratesneuralnetworkscurseofdimensionalityminimaxHammersley-Cliffordfactorizationgraphcliquesize
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

This paper tries to show that the curse of dimensionality in nonparametric density estimation disappears when the target density is Markov to a known graph: the convergence rate is governed by the largest clique $r$ of that graph, not by the ambient dimension $d$. It proves that a ReLU-network estimator trained by the squared-error objective attains $\|p-\hat p_n\|_1 \in \widetilde{O}_p(n^{-1/(4+r)})$ for every positive Lipschitz density satisfying the Markov property, and it proves the minimax $L^1$ rate for the same class is $\widetilde{O}_p(n^{-1/(2+r)})$, so $r$ is the effective dimension. Under power path and grid graph models the paper argues that images, audio, video, and text have $r=O(1)$, giving for example an effective dimension of $9$ for a $32\times 32$ image instead of $1024$. A reader should care because, if the claim holds, it provides a structural explanation for why deep networks can learn very high-dimensional densities from modest sample sizes, complementing the manifold hypothesis.

What carries the argument

The load-bearing object is the Hammersley-Clifford factorization of a positive MRF density into a product of functions on maximal cliques, together with ReLU-network approximation and empirical $L^2$ minimization over the product architecture $F^* = \{\prod_{V'\in C(G)} q_{V'}\circ e_{V'}\}$. Proposition A.1 shows the clique factors are Lipschitz and bounded, Schmidt-Hieber's approximation theorem supplies a network for each factor whose size depends on $r$ rather than $d$, and a covering-number bound controls the empirical process uniformly over $F^*$. The near-optimal estimator replaces networks by histograms on clique blocks and selects among them with a Scheffé tournament, a minimum-distance choice over a finite cover of candidate densities.

What would settle it

Take $p$ on $[0,1]^d$ to be a product of $d$ independent copies of a fixed positive Lipschitz one-dimensional marginal, which is Markov to the empty graph with $r=1$; run the paper's neural estimator with that graph known on $n$ samples for $d=2,10,100$. If the log-log slope of $L^1$ error against $n$ moves away from about $-1/5$ as $d$ grows, the claimed dimension independence fails.

Watch

Extended reading notes

Core claim

The central claim is Theorem 4.2: for any finite graph $G$ with largest clique size $r$, there is a known sequence of neural network architectures $F^*$ such that $\hat p_n = \arg\min_{f\in F^*}(\|f\|_2^2 - (2/n)\sum_{i=1}^n f(x_i))$ satisfies $\|p-\hat p_n\|_1 \in \widetilde{O}_p(n^{-1/(4+r)})$ whenever $p$ is a positive Lipschitz density Markov to $G$. Theorem 4.8 adds a computationally intractable Scheffé-tournament estimator that attains the minimax $L^1$ rate $\widetilde{O}_p(n^{-1/(2+r)})$, and the paper's lower-bound argument shows no estimator can do better, so the effective statistical dimension is exactly $r$. The argument runs through the Hammersley-Clifford factorization $p(x)=\prod_{V'\in C(G)}\psi_{V'}(x_{V'})$, which reduces approximation of a $d$-dimensional density to approximation of factors depending on at most $r$ coordinates each.

Load-bearing premise

The rates require the estimator to be handed the graph $G$ and its maximal cliques, and the practical claims require real images, audio, video, and text to be exactly Markov to such graphs with $r=O(1)$; if the graph is unknown or the conditional independencies hold only approximately, the stated dimension-independent rates do not follow.

Editorial extensions

If this is right

  • For image models, Corollary 4.5 gives $\widetilde{O}_p(n^{-1/7})$ for the squared grid graph $L^2_{d\times d'}$ and $\widetilde{O}_p(n^{-1/9})$ for the diagonal grid $(L^+_{d\times d'})^2$, with no dependence on the number of pixels.
  • For sequential data under the $t$-th power of a path graph, the largest clique has size $t+1$, so longer-range dependence worsens the rate only polynomially in $t$, not exponentially in sequence length.
  • The minimax bound of Theorem 4.8 implies the dimension-free behavior is intrinsic to the problem class: no estimator can beat the $n^{-1/(2+r)}$ rate on all such densities.
  • For tree-structured densities ($r=2$), the neural estimator runs at $\widetilde{O}_p(n^{-1/6})$ while the optimal rate is $\widetilde{O}_p(n^{-1/4})$, quantifying the gap between the tractable squared-loss network and the intractable optimal estimator.

Reading between the lines

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

  • A direct next step is to estimate the graph or its clique size from data; the paper fixes $G$ and constructs $F^*$ from $C(G)$, but the rate formulas suggest that model selection over graph powers with a penalty proportional to $r$ could yield an adaptive, still dimension-free estimator.
  • The same clique-factorization proof strategy should transfer to other losses, such as negative log-likelihood, whenever the required empirical-process and approximation bounds hold; the paper analyzes the squared loss and the tournament estimator, leaving other objectives open.
  • Real data are at best approximately Markov, so one expects the clean exponent to acquire an additive term measuring the distance from the true density to the nearest exactly-Markov density; testing this would require conditional-independence checks beyond the paper's pairwise scatter plots.
  • If both structures are present, the effective dimension of a problem could plausibly be $\min(\text{intrinsic manifold dimension}, r)$, unifying the paper's independence-based explanation with the manifold hypothesis.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes that for densities satisfying the Markov property with respect to a graph G whose maximum clique has size r, a neural-network estimator minimizing an L2-type loss converges in L1 at the dimension-independent rate eO_p(n^{-1/(4+r)}), and that the minimax L1 rate is eO_p(n^{-1/(2+r)}). The authors model images, audio, video, and text as Markov random fields on power graphs, arguing that the maximum clique size is O(1) in practice, and provide scatter-plot evidence for image data. The proof of the main upper bound (Theorem 4.2) is in Appendix B; the near-optimal estimator and lower bound are in Appendix C and E.

Significance. If the main theorem were fully established, the paper would offer a genuinely new explanation for deep learning's success in high-dimensional density estimation, complementary to the manifold hypothesis: the effective dimension would be the maximum clique size r rather than the ambient dimension d. The paper has several strengths: the graph-theoretic clique bounds (Lemmas 4.3, 4.4, 4.6) are explicit and checkable; the lower-bound reduction in Appendix E is standard and sound; and the paper is honest that the near-optimal estimator in Theorem 4.8 is computationally intractable. However, the central upper-bound theorem currently has a load-bearing gap: the architecture F* in the proof depends on the unknown density through a constant C chosen to exceed the Hammersley-Clifford factor bounds, so Theorem 4.2 as stated is not proven for a density-independent architecture. The practical significance claims are also weakened by the fact that the analyzed estimator uses the exact L2 norm of the network rather than the stochastic objective described in Section 4.1.

major comments (4)
  1. [Appendix B (proof of Theorem 4.2), eq. (9)] The architecture F* constructed in the proof is not determined by G and n alone. In Appendix B the proof states 'We pick a sufficiently large C that is greater than Cψ' and then defines FV' = F(ℓV', wV', s, C), where Cψ is the uniform upper bound on the Hammersley-Clifford factors of the unknown density p. Since F* in eq. (9) is built from these FV', the architecture depends on p. The theorem statement's quantifier ('There exists a known sequence of architectures F* ... for any ... p') requires F* to be independent of p. No finite C works uniformly: for the empty graph on two vertices, the density p_α(x,y) = (α/(e^α-1))^2 e^{α(x+y)} has factor ψ_α(x) = α e^{αx}/(e^α-1) with sup ≈ α, so any fixed C is exceeded for large α. Thus Theorem 4.2 is either an oracle statement with F* depending on p, or the stated result is unproven. Because this is the paper's central dimension-independent claim, this gap is load-bearing.
  2. [Theorem 4.2 and Section 4.1] The estimator analyzed in Theorem 4.2 is p_hat_n = argmin_{f∈F*} (||f||_2^2 - (2/n)Σ_i f(x_i)), which requires the exact L2 norm of every candidate network on [0,1]^d. This is not the 'simple L2-minimizing loss' claimed in the abstract, and it is not the stochastic objective introduced in Section 4.1, where the integral of f^2 is replaced by a Monte Carlo average over uniform samples. No theorem is stated or proved for the Monte Carlo version, and no argument is given that the exact L2 norm of a ReLU network can be computed efficiently. The practical relevance of the rate therefore rests on an analysis that is not in the paper.
  3. [Eq. (9), Section 4.2] The class F* is a set of products of ReLU networks qV'∘eV'. The paper refers to this as a 'neural network architecture,' but a product of networks is not itself a ReLU network unless multiplication can be represented exactly by ReLU layers of the stated size. The proof uses the product structure directly (Lemma B.2) and never constructs a single ReLU network realizing F*. If products are only approximated, the exact-argmin analysis does not apply. The manuscript should either define the estimator as a product network and adjust the terminology, or prove that F* is a standard ReLU architecture.
  4. [Section 3, Figure 3] The empirical support for the MRF assumption in image data is based on scatter plots of pixel values and reported reductions in correlation after conditioning on one adjacent pixel. These plots do not test the conditional independence required by the model: the power-grid MRF requires pixels such as (8,8) and (9,12) to be independent given the entire separating set, not given a single pixel, and correlation is only one measure of dependence. The sentence 'This provides strong evidence for the validity of the MRF model' (Section 3) overstates what the experiments show. The numerical claims in Section 4.3 about CIFAR-10's effective dimension assume the exact MRF model, so this evidence is load-bearing for the paper's application claims.
minor comments (5)
  1. [Appendix D] The displayed edge conditions for L_{d×d'} and L^+_{d×d'} use |i-j|+|i'-j'| and max{|i-j|,|i'-j'|}; these should be |i-i'|+|j-j'| and max{|i-i'|,|j-j'|}. Similarly, the definition of L^t_d should use |i-j| ≤ t, not |i-j| ≥ t.
  2. [Section 4.4] Theorem 4.8 is stated under a 'strong density assumption' that is not defined in the main text; the formal statement in Appendix C assumes a positive Lipschitz density. Please reconcile the two statements.
  3. [Appendix C, eq. (26)] The displayed equality min_q ∥p-q∥_1 = min_q ∥∏(ψ_{V'}∘e_{V'} - q_{V'})∥_1 is not correct; the argument should be applied to ∏ψ_{V'} - ∏q_{V'} directly via Lemma C.2.
  4. [Section 4.1] The sentence 'The weights are constrained to [-1,1], effectively implementing weight decay via constrained optimization' is misleading: bounded weights are a projection, not weight decay.
  5. [Section 4.3] There is a typo, 'exmaple', in the sentence 'Judging from the exmaple in Figure 3'. Also, Lemmas 4.3 and 4.4 should state the condition on t as t < min(d,d') rather than t < d,d'.

Circularity Check

2 steps flagged · score 6.0 of 10

Theorem 4.2's architecture is chosen using the unknown target's Hammersley-Clifford factor bound, so the main rate is established only for an oracle class, not for the claimed known sequence of architectures.

  1. self definitional [Appendix B, proof of Theorem B.1 (used for Theorem 4.2), around eq. (9)]
    "For any V ′ ∈ C(G), let FV ′ = F (ℓV ′, wV ′, s, C) where ℓV ′, wV ′, s, C will be determined later. ... We pick a sufficiently large C that is greater than Cψ."

    Cψ is introduced in Proposition A.1 as a bound on the unknown Hammersley-Clifford factors ψV′ of p; it is not determined by G or n. Choosing C > Cψ means F*, the hypothesis class minimized by p̂n, is built using knowledge of the target p. The theorem claims a known sequence of architectures valid for every positive Lipschitz p Markov to G, but the proof supplies only a p-dependent class. For the empty graph on two variables, p_α(x,y) ∝ e^{α(x+y)} has singleton factors with sup ≈ α, so no fixed C can cover all such p with r = 1; the n^{-1/(4+r)} rate is therefore established only for an oracle estimator whose class has been tuned to p. The same p-dependence enters through Lψ as K in Schmidt-Hieber's Theorem B.3.

  2. self definitional [Appendix C, proof of Theorem C.1 (Theorem 4.8), after eq. (25)]
    "Because ψV ′ are all Lipschitz continuous on a bounded set, they must all be bounded and, for sufficiently large n, ψV ′ ≤ C for all V ′ ∈ C(G)."

    Here again C defines Qd,b,V′,C and hence Q_n and eQ_n, the collection over which the Scheffé tournament estimates. The bound ψV′ ≤ C is a property of the unknown target p (Proposition A.1), not of G and n, so the estimator V_n is defined using the very density it estimates. This is the same self-definitional structure as in Theorem 4.2, though the minimax lower bound and tournament argument themselves are standard.

full rationale

The rate calculation is otherwise self-contained: Proposition A.1 is an external Hammersley-Clifford factorization (via Chang's notes), the approximation bound is Schmidt-Hieber's external ReLU network theorem, the empirical-process term is a standard Hoeffding/covering argument, and the lower bound reduces to the standard r-dimensional Lipschitz minimax rate. No data are fitted and no constants are calibrated to a subset of the target. The circularity is concentrated in the choice of F*: C and Lψ come from the unknown p, so the estimator is an oracle. The same p-dependent C recurs in the histogram tournament of Theorem 4.8. I therefore score 6: the central claim partially reduces, by construction, to knowing the target's factorization bounds; it is not a case of self-citation or fitted-input prediction.

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

The theoretical results rest on standard theorems plus the structural domain assumption that the density is a positive Lipschitz MRF with known graph. The empirical applicability to real data additionally assumes exact MRF structure with constant clique size, which is supported only informally.

assumptions (5)
  • standard math Hammersley-Clifford theorem: a positive density satisfying the Markov property with respect to G factorizes as a product of functions on the maximal cliques.
    Invoked in Proposition A.1 and Section 4.1; the positivity of p is required for this factorization.
  • standard math Schmidt-Hieber's ReLU network approximation theorem (Theorem B.3) and the associated covering number bound (Lemma B.4).
    Used to bound approximation error of the clique factors by neural networks in the proof of Theorem 4.2.
  • standard math Yatracos/Scheffé tournament selection algorithm (Theorem C.7 from Ashtiani et al. and Devroye-Lugosi) for choosing among a finite collection of densities.
    Used in the proof of the optimal L1 rate in Theorem 4.8.
  • domain assumption The target density is positive and Lipschitz continuous on [0,1]^d, and is Markov with respect to a known graph G with known maximal cliques.
    Stated in Theorem 4.2 and Proposition A.1; positivity is needed for Hammersley-Clifford and boundedness below.
  • domain assumption Real-world data (images, audio, video, text) are approximately Markov to power graphs with constant maximum clique size r = O(1).
    Argued in Section 3 using correlation scatter plots (Figure 3); no statistical test of conditional independence is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dimension-independent rates for structured neural density estimation." pith.science (2026). https://pith.science/paper/D47M4HZI

@misc{pith2026241115095,
  author       = {Pith},
  title        = {Pith review of: Dimension-independent rates for structured neural density estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D47M4HZI}},
  note         = {Machine review of arXiv:2411.15095}
}
abstract

We show that deep neural networks achieve dimension-independent rates of convergence for learning structured densities such as those arising in image, audio, video, and text applications. More precisely, we demonstrate that neural networks with a simple $L^2$-minimizing loss achieve a rate of $n^{-1/(4+r)}$ in nonparametric density estimation when the underlying density is Markov to a graph whose maximum clique size is at most $r$, and we provide evidence that in the aforementioned applications, this size is typically constant, i.e., $r=O(1)$. We then establish that the optimal rate in $L^1$ is $n^{-1/(2+r)}$ which, compared to the standard nonparametric rate of $n^{-1/(2+d)}$, reveals that the effective dimension of such problems is the size of the largest clique in the Markov random field. These rates are independent of the data's ambient dimension, making them applicable to realistic models of image, sound, video, and text data. Our results provide a novel justification for deep learning's ability to circumvent the curse of dimensionality, demonstrating dimension-independent convergence rates in these contexts.

Figures

Figures reproduced from arXiv: 2411.15095 by the authors.

Figure 1
Figure 1. (a): Examples of conditioning X = x (dotted lines) when a density’s support is a man￾ifold (solid lines). (b): Highly simplified examples of common MRF graphs. Paths correspond to sequential data and grids to spatial. In this work, we show that, for a very general class of densities that are Markov to an undirected graph (a.k.a. MRF), density estimation can be achieved with a neural network and a simple L 2 minimizi… view at source ↗
Figure 3
Figure 3. Top row: Scatterplots comparing the grayscale values of pixel (8,8) with various other pixels for 100 randomly selected images. The decreasing correlation between pixels as their distance increases is evident. Bottom row: The same comparisons as the top row, but conditioned on pixel (9,8) having a value approximately equal to 0.48 (the median value for this pixel across the dataset). Note the increased concentration… view at source ↗
Figure 2
Figure 2. An example MRF. The random variables x and y are independent given z. One of the most well-known MRFs is the Markov chain. A Markov chain of length d corresponds to the “path” graph Ld of d random variables. The above example with x, y, z corresponds to the graph L3 and the MRF corresponding to L4 shown in Figure 1b. In a Markov chain, the indices are often interpreted as a time parameter. A classic example is a gam… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: The leftmost image (a) is a 640 × 427 pixel photograph from the COCO 2014 dataset (Lin et al., 2014). Image (b) shows an enlarged version of the 102 × 102 pixel region outlined in (a). Images (c) and (d) display the 12-pixel and 1-pixel width borders of that region, re…
Figure 5
Figure 5. Figure 5: Illustrations of a path graph and its powers. Left: The path graph [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Comparison of vertex neighborhoods in different graph structures. (a) Neighborhood of [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: A graph with max￾imal cliques denoted by sur￾rounding rectangles. For neural networks we investigate estimators of the form: pˆ(x) = Y V ′∈C(G) ψbV ′ (xV ′ ), where ψbV ′ are ReLU networks with architectures dependent only on G, |V ′ |, and the sample size n. The weigh…
Figure 8
Figure 8. Figure 8: This figure presents scatter plots analogous to those in Figure [PITH_FULL_IMAGE:figures/full_fig_p027_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 31 canonical work pages

  1. [1]

    Nearly tight sample complexity bounds for learning mixtures of gaussians via sample compression schemes

    Hassan Ashtiani, Shai Ben-David, Nicholas Harvey, Christopher Liaw, Abbas Mehrabian, and Yaniv Plan. Nearly tight sample complexity bounds for learning mixtures of gaussians via sample compression schemes. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett (eds.), Advances in Neural Information Processing Systems 31, pp.\...

    work page 2018

  2. [2]

    Universal approximation bounds for superpositions of a sigmoidal function

    Andrew R Barron. Universal approximation bounds for superpositions of a sigmoidal function. IEEE Transactions on Information theory, 39 0 (3): 0 930--945, 1993

    1993

  3. [3]

    Representation learning: A review and new perspectives

    Yoshua Bengio, Aaron Courville, and Pascal Vincent. Representation learning: A review and new perspectives. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35 0 (8): 0 1798--1828, 2013. doi:10.1109/TPAMI.2013.50

    doi:10.1109/tpami.2013.50 2013

  4. [4]

    Estimating a density near an unknown manifold: a Bayesian nonparametric approach

    Cl \'e ment Berenfeld, Paul Rosa, and Judith Rousseau. Estimating a density near an unknown manifold: a bayesian nonparametric approach. arXiv preprint arXiv:2205.15717, 2022

    work page Pith review arXiv 2022

  5. [5]

    Why deep learning works: A manifold disentanglement perspective

    Pratik Prabhanjan Brahma, Dapeng Wu, and Yiyuan She. Why deep learning works: A manifold disentanglement perspective. IEEE Transactions on Neural Networks and Learning Systems, 27 0 (10): 0 1997--2008, 2016. doi:10.1109/TNNLS.2015.2496947

    arXiv 1997

  6. [6]

    H. Cao, C. Tan, Z. Gao, Y. Xu, G. Chen, P. Heng, and S. Z. Li. A survey on generative diffusion models. IEEE Transactions on Knowledge and Data Engineering, 36 0 (07): 0 2814--2830, jul 2024. ISSN 1558-2191. doi:10.1109/TKDE.2024.3361474

    arXiv 2024

  7. [7]

    On the Local Behavior of Spaces of Natural Images

    Gunnar Carlsson, Tigran Ishkhanov, Vin de Silva, and Afra Zomorodian. On the Local Behavior of Spaces of Natural Images . International Journal of Computer Vision, 76 0 (1): 0 1--12, January 2008. ISSN 1573-1405. doi:10.1007/s11263-007-0056-x. URL https://doi.org/10.1007/s11263-007-0056-x

    work page doi:10.1007/s11263-007-0056-x 2008

  8. [8]

    Stochastic processes, 2007

    Joseph Chang. Stochastic processes, 2007. URL http://www.stat.yale.edu/ pollard/Courses/251.spring2013/Handouts/Chang-notes.pdf. Accessed: Oct 1, 2024

    work page 2007

Show all 50 references
  1. [9]

    Structured neural networks for density estimation and causal inference

    Asic Chen, Ruian Ian Shi, Xiang Gao, Ricardo Baptista, and Rahul G Krishnan. Structured neural networks for density estimation and causal inference. Advances in Neural Information Processing Systems, 36, 2024

    2024

  2. [10]

    Score-based generative models break the curse of dimensionality in learning a family of sub-gaussian probability distributions

    Frank Cole and Yulong Lu. Score-based generative models break the curse of dimensionality in learning a family of sub-gaussian probability distributions. arXiv preprint arXiv:2402.08082, 2024

    2024 arXiv

  3. [11]

    Devroye and L

    L. Devroye and L. Gyorfi. Nonparametric Density Estimation: The L^1 View. Wiley Interscience Series in Discrete Mathematics. Wiley, 1985. ISBN 9780471816461

    1985

  4. [12]

    Devroye and G

    L. Devroye and G. Lugosi. Combinatorial Methods in Density Estimation. Springer, New York, 2001

    2001

  5. [13]

    Gerald B. Folland. Real analysis: modern techniques and their applications. Pure and applied mathematics. Wiley, 1999. ISBN 9780471317166

    1999

  6. [14]

    Made: Masked autoencoder for distribution estimation

    Mathieu Germain, Karol Gregor, Iain Murray, and Hugo Larochelle. Made: Masked autoencoder for distribution estimation. In International Conference on Machine Learning, pp.\ 881--889. PMLR, 2015

    2015

  7. [15]

    Tree density estimation

    L \'a szl \'o Gy \"o rfi, Aryeh Kontorovich, and Roi Weiss. Tree density estimation. IEEE Transactions on Information Theory, 69 0 (2): 0 1168--1176, 2022

    2022

  8. [16]

    Markov fields on finite graphs and lattices

    John M Hammersley and Peter Clifford. Markov fields on finite graphs and lattices. Unpublished manuscript, 46, 1971

    1971

  9. [17]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin (eds.), Advances in Neural Information Processing Systems, volume 33, pp.\ 6840--6851. Curran Associates, Inc., 2020. URL http...

    2020

  10. [18]

    Uniform convergence rates for kernel density estimation

    Heinrich Jiang. Uniform convergence rates for kernel density estimation. In International Conference on Machine Learning, pp.\ 1694--1703, 2017

    2017

  11. [19]

    Deep nonparametric regression on approximate manifolds: Nonasymptotic error bounds with polynomial prefactors

    Yuling Jiao, Guohao Shen, Yuanyuan Lin, and Jian Huang. Deep nonparametric regression on approximate manifolds: Nonasymptotic error bounds with polynomial prefactors. The Annals of Statistics, 51 0 (2): 0 691--716, 2023

    2023

  12. [20]

    Composing graphical models with neural networks for structured representations and fast inference

    Matthew J Johnson, David K Duvenaud, Alex Wiltschko, Ryan P Adams, and Sandeep R Datta. Composing graphical models with neural networks for structured representations and fast inference. In D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett (eds.), Advances in Neu...

    2016

  13. [21]

    Robert W. Keener. Bayesian Inference: Modeling and Computation, pp.\ 301--318. Springer New York, New York, NY, 2010. ISBN 978-0-387-93839-4. doi:10.1007/978-0-387-93839-4_15. URL https://doi.org/10.1007/978-0-387-93839-4_15

    2010 doi

  14. [22]

    Causal autoregressive flows

    Ilyes Khemakhem, Ricardo Monti, Robert Leech, and Aapo Hyvarinen. Causal autoregressive flows. In International Conference on Artificial Intelligence and statistics, pp.\ 3520--3528. PMLR, 2021

    2021

  15. [23]

    Klusowski and Andrew R

    Jason M. Klusowski and Andrew R. Barron. Approximation by combinations of relu and squared relu ridge functions with ^1 and ^0 controls. IEEE Transactions on Information Theory, 64 0 (12): 0 7649--7656, 2018. doi:10.1109/TIT.2018.2874447

    2018

  16. [24]

    Minimax optimal density estimation using a shallow generative model with a one-dimensional latent variable

    Hyeok Kyu Kwon and Minwoo Chae. Minimax optimal density estimation using a shallow generative model with a one-dimensional latent variable. In International Conference on Artificial Intelligence and Statistics, pp.\ 469--477. PMLR, 2024

    2024

  17. [25]

    S. Z. Li. Markov random field models in computer vision. In Jan-Olof Eklundh (ed.), Computer Vision --- ECCV '94, pp.\ 361--370, Berlin, Heidelberg, 1994. Springer Berlin Heidelberg. ISBN 978-3-540-48400-4

    1994

  18. [26]

    How well can generative adversarial networks ( GAN ) learn densities: A nonparametric view

    Tengyuan Liang. How well can generative adversarial networks ( GAN ) learn densities: A nonparametric view. arXiv preprint arXiv:1712.08244, 2017

    2017 arXiv

  19. [27]

    Lawrence Zitnick , and Piotr Doll \'a r

    Tsung-Yi Lin , Michael Maire , Serge Belongie , Lubomir Bourdev , Ross Girshick , James Hays , Pietro Perona , Deva Ramanan , C. Lawrence Zitnick , and Piotr Doll \'a r . Microsoft COCO: Common Objects in Context . arXiv e-prints, art. arXiv:1405.0312, May 2014. doi:10.48550/a...

    2014 doi arXiv

  20. [28]

    Forest density estimation

    Han Liu, Min Xu, Haijie Gu, Anupam Gupta, John Lafferty, and Larry Wasserman. Forest density estimation. The Journal of Machine Learning Research, 12: 0 907--951, 2011

    2011

  21. [29]

    The barron space and the flow-induced function spaces for neural network models

    Chao Ma, Lei Wu, et al. The barron space and the flow-induced function spaces for neural network models. Constructive Approximation, 55 0 (1): 0 369--406, 2022

    2022

  22. [30]

    Neural networks for density estimation

    Malik Magdon-Ismail and Amir Atiya. Neural networks for density estimation. Advances in Neural Information Processing Systems, 11, 1998

    1998

  23. [31]

    Adaptive approximation and generalization of deep neural network with intrinsic dimensionality

    Ryumei Nakada and Masaaki Imaizumi. Adaptive approximation and generalization of deep neural network with intrinsic dimensionality. Journal of Machine Learning Research, 21 0 (174): 0 1--38, 2020

    2020

  24. [32]

    Diffusion models are minimax optimal distribution estimators

    Kazusato Oko, Shunta Akiyama, and Taiji Suzuki. Diffusion models are minimax optimal distribution estimators. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan Scarlett (eds.), Proceedings of the 40th International Conference on M...

    2023

  25. [33]

    Deep generative models: Survey

    Achraf Oussidi and Azeddine Elhassouny. Deep generative models: Survey. In 2018 International Conference on Intelligent Systems and Computer Vision (ISCV), pp.\ 1--8, 2018. doi:10.1109/ISACV.2018.8354080

    2018

  26. [34]

    Submanifold density estimation

    Arkadas Ozakin and Alexander Gray. Submanifold density estimation. In Y. Bengio, D. Schuurmans, J. Lafferty, C. Williams, and A. Culotta (eds.), Advances in Neural Information Processing Systems, volume 22. Curran Associates, Inc., 2009. URL https://proceedings.neurips.cc/pape...

    2009

  27. [35]

    Kernel density estimation on riemannian manifolds

    Bruno Pelletier. Kernel density estimation on riemannian manifolds. Statistics & probability letters, 73 0 (3): 0 297--304, 2005

    2005

  28. [36]

    The intrinsic dimension of images and its impact on learning

    Phil Pope, Chen Zhu, Ahmed Abdelkader, Micah Goldblum, and Tom Goldstein. The intrinsic dimension of images and its impact on learning. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=XJk19XzGq2J

    2021

  29. [37]

    A Useful Convergence Theorem for Probability Distributions

    Henry Scheffe. A Useful Convergence Theorem for Probability Distributions . The Annals of Mathematical Statistics, 18 0 (3): 0 434 -- 438, 1947. doi:10.1214/aoms/1177730390. URL https://doi.org/10.1214/aoms/1177730390

    1947

  30. [38]

    Nonparametric regression using deep neural networks with relu activation function

    Johannes Schmidt-Hieber. Nonparametric regression using deep neural networks with relu activation function. arXiv preprint arXiv:1708.06633, 2017

    2017 arXiv

  31. [39]

    Deep ReLU network approximation of functions on a manifold

    Johannes Schmidt-Hieber. Deep ReLU network approximation of functions on a manifold. arXiv preprint arXiv:1908.00695, 2019

    1908 arXiv

  32. [40]

    Nonparametric density estimation under adversarial losses

    Shashank Singh, Ananya Uppal, Boyue Li, Chun-Liang Li, Manzil Zaheer, and Barnab \'a s P \'o czos. Nonparametric density estimation under adversarial losses. Advances in Neural Information Processing Systems, 31, 2018

    2018

  33. [41]

    Adaptivity of diffusion models to manifold structures

    Rong Tang and Yun Yang. Adaptivity of diffusion models to manifold structures. In International Conference on Artificial Intelligence and Statistics, pp.\ 1648--1656. PMLR, 2024

    2024

  34. [42]

    Tsybakov

    A.B. Tsybakov. Introduction to nonparametric estimation. Springer Series in Statistics, New York, pp.\ 214, 2009. cited By 1

    2009

  35. [43]

    Nonparametric density estimation & convergence rates for GAN s under B esov IPM losses

    Ananya Uppal, Shashank Singh, and Barnab \' a s P \' o czos. Nonparametric density estimation & convergence rates for GAN s under B esov IPM losses. In Hanna M. Wallach, Hugo Larochelle, Alina Beygelzimer, Florence d'Alch \' e - Buc, Emily B. Fox, and Roman Garnett (eds.), Adv...

    2019

  36. [44]

    Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance

    Jonathan Weed and Francis Bach. Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance . Bernoulli, 25 0 (4A): 0 2620 -- 2648, 2019. doi:10.3150/18-BEJ1065. URL https://doi.org/10.3150/18-BEJ1065

    2019 doi

  37. [45]

    Graphical normalizing flows

    Antoine Wehenkel and Gilles Louppe. Graphical normalizing flows. In International Conference on Artificial Intelligence and Statistics, pp.\ 37--45. PMLR, 2021

    2021

  38. [46]

    Yatracos

    Yannis G. Yatracos. Rates of Convergence of Minimum Distance Estimators and Kolmogorov's Entropy . The Annals of Statistics, 13 0 (2): 0 768 -- 774, 1985. doi:10.1214/aos/1176349553. URL https://doi.org/10.1214/aos/1176349553

    1985

  39. [48]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  40. [49]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  41. [50]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  42. [51]

    Minimax Distribution Estimation in W asserstein Distance

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2020 arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.