Pith. sign in

REVIEW 4 major objections 5 minor 58 references

Time-Varying Graph Learning for Data with Heavy-Tailed Distribution

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read Heavy-tailed graph learner clusters stocks at 69% accuracy

desk verdict A genuinely useful robust time-varying graph learner whose probabilistic foundation has a rank-subtlety bug in Eq. (10) and whose experiments need more discipline; worth a serious referee. read the letter →

arxiv 2501.00606 v1 pith:UC3IF5M6 submitted 2024-12-31 cs.LG

classification cs.LG MSC 62H3005C5062F15
keywords time-varyinggraphlearningheavy-taileddistributionStudent-tk-componentspectralclusteringLaplacianADMMfinancialdata
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 that a time-varying graph can be learned reliably from heavy-tailed, noisy, partially observed signals by solving a single maximum a posteriori problem, and that the learned graph is directly usable for clustering. The proposed model treats each frame's signal as multivariate Student-t whose scale is set by the graph Laplacian, lets the edge weights evolve through a non-negative vector autoregression, and constrains the Laplacian to have exactly $p-k$ nonzero eigenvalues so the graph has $k$ connected components. If this is right, dynamic graph estimation and spectral clustering become one operation instead of two, and financial applications inherit a method that tolerates outliers without switching to a different model family. The paper's evidence is synthetic F-scores and real S&P 500 experiments, where clustering accuracy on the final frame reaches 0.69 and a graph-smoothness portfolio backtests to a Sharpe ratio of 1.31.

What carries the argument

The carrying object is the Student-t-Laplacian signal model of Eq. (10), combined with the rank-deficiency constraint $\operatorname{rank}(L_w)=p-k$ that forces the learned graph to have exactly $k$ connected components. The Laplacian $L_w$ plays the role of a precision-like matrix in the Student-t density, and the generalized determinant $\det_*$ is the product of its nonzero eigenvalues, so the likelihood stays well defined when the Laplacian is singular because of the $k$ components. A non-negative VAR equation $w_n=(a\odot w_{n-1}+\epsilon_n)_+$ with Laplace innovations provides the temporal prior on edge weights, including a multiplicative coefficient vector $a$ instead of the usual additive smoothness penalty. The solver is an ADMM loop whose subproblems use majorization-minimization to get closed-form updates for the graph weights, the signals, and the VAR coefficients, and the theory shows the augmented Lagrangian converges and the constraints are satisfied at the limit.

What would settle it

Generate synthetic signals from a multivariate Student-t whose scale matrix is not the pseudo-inverse of a graph Laplacian, such as a general sparse precision matrix, and feed noisy, partially observed versions through Algorithm 1; if the learned $k$-component graph and clustering accuracy do not clearly beat the same benchmarks used in the paper, the claim that the Student-t-Laplacian model is the right heavy-tailed dependence structure fails. A second check is to run the same experiment with vertex degrees that vary across frames, where the fixed constraint $d_w=d$ is violated, and measure how much the method's clustering accuracy degrades.

Watch

Extended reading notes

Core claim

The central claim is that solving the MAP problem of Eq. (13) with the proposed ADMM/MM algorithm recovers the true time-varying $k$-component graph from Student-t distributed observations even with missing entries and additive Gaussian noise, and that the recovered graph beats existing time-varying graph learners on heavy-tailed data. The key identity is the Student-t likelihood $p(x_t\mid w_n)\propto \det_*(L_{w_n})^{1/2}(1+x_t^\top L_{w_n}x_t/\nu)^{-(\nu+p)/2}$, which makes the graph Laplacian the precision-like object while the generalized determinant $\det_*$ (product of nonzero eigenvalues) handles the rank-deficient $k$-component Laplacian. The paper argues that the rank constraint $\operatorname{rank}(L_w)=p-k$ plus the degree constraint $d_w=d$ yields a graph whose connected components are the desired clusters, and that the non-negative VAR prior $w_n=(a\odot w_{n-1}+\epsilon_n)_+$ correctly captures the graph's temporal evolution. Convergence of the iterative solver is proven, and on S&P 500 log-returns the method reaches ACC 0.69, ARI 0.58, and modularity 0.62 on the final frame, with a backtested Sharpe ratio of 1.31 versus 1.00 for the maximum Sharpe ratio portfolio.

Load-bearing premise

The estimator is only as trustworthy as the generative model: each frame's signal really is multivariate Student-t with scale set by the graph Laplacian, the noise is i.i.d. Gaussian with known variance, the graph is piecewise constant, and the vertex degrees are fixed at a known constant $d$; if real heavy-tailed data have a different dependence structure, the MAP objective is misspecified and the reported gains may not transfer.

Editorial extensions

If this is right

  • Graph updates need only one mini-batch frame, so the method runs in a semi-online mode that balances delay and statistical accuracy by choosing the frame length $T_n$.
  • The learned graph has exactly $k$ connected components, so clustering is read directly off the graph rather than requiring a separate spectral-clustering stage after learning.
  • For large $\nu$ the Student-t likelihood approaches the Gaussian case, making the estimator a drop-in replacement for existing Gaussian time-varying graph learners while still tolerating outliers.
  • On real S&P 500 data the estimator separates 100 stocks into 8 sectors with ACC 0.69 and ARI 0.58 on the final frame, beating both static heavy-tailed graph learners and online time-varying baselines.
  • Portfolio weights chosen to maximize the ratio of expected return over graph smoothness backtest to a Sharpe ratio of 1.31, compared with 1.00 for the maximum Sharpe ratio portfolio.

Reading between the lines

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

  • Beyond the paper: a natural ablation is to run Algorithm 1 on clean Gaussian data, where the Student-t machinery should reduce to the Gaussian case; the comparison would show how much of the gain comes from heavy tails versus from the rank and $k$-component constraint alone.
  • Beyond the paper: the fixed-degree constraint $d_w=d$, set to 1 in all experiments, is a strong structural prior that real networks rarely satisfy, so allowing estimated or time-varying degrees is a testable extension that would reveal how much of the clustering gain depends on this assumption.
  • Beyond the paper: the multiplicative coefficient vector $a$ in the VAR prior is interpretable as per-edge persistence, and one could test whether the learned $a$ identifies edges that persist across market regimes, which is a falsifiable claim about financial network dynamics.
  • Beyond the paper: the portfolio construction still relies on external robust estimates of mean and covariance, so an end-to-end version using only the learned Laplacian as the risk object would clarify whether the graph alone carries the portfolio signal.
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.

Referee Report

4 major / 5 minor

Summary. The paper proposes a semi-online (mini-batch) method for learning time-varying undirected graphs from heavy-tailed data. The signal in each frame is modeled by a rank-deficient Student-t distribution tied to the graph Laplacian, the graph weights evolve according to a non-negative VAR model, and the optimization is solved by ADMM with majorization-minimization inner steps. The method enforces k-component structure through rank and fixed-degree constraints and is evaluated on synthetic data and S&P 500 data for clustering and portfolio design.

Significance. If the method works as claimed, it fills a real gap: existing time-varying graph learning methods are mostly Gaussian and do not produce k-component graphs useful for clustering. The paper contributes an explicit ADMM/MM algorithm with closed-form updates, handles missing data and noise, and reports substantial F-score improvements on synthetic data (e.g., 0.60 vs. 0.40 at SR=1) as well as strong clustering metrics (ACC 0.69, ARI 0.58) on financial data. The temporal VAR model for graph weights is a useful addition. However, the probabilistic foundation has a rank-exponent mismatch, and the convergence proof rests on a non-constructive condition; these issues must be resolved before the claims can be accepted.

major comments (4)
  1. [Section III, Eq. (10) and Appendix A] The likelihood in Eq. (10) is not a valid density for the stated generative model when the graph has k components and rank r=p-k. For x_t = (L†)^{1/2} ν_t with ν_t ~ St_ν(0,I_p), the marginal density is proportional to det*(L)^{1/2} (1 + x^T L x / ν)^{-(ν+r)/2}, not (ν+p)/2. The paper's Eq. (13) and Proposition 1 consequently do not follow as MAP from the model in Eq. (10), and the synthetic experiments in Section IV-A (which generate data exactly as above) use a likelihood with the wrong exponent. The coefficient (ν+p)/T_n in Eq. (13) should be (ν+r)/T_n, and this correction propagates to the majorizers and updates in Section III. Please correct the exponent throughout and re-derive the algorithm, or provide a different justification for using (ν+p)/2.
  2. [Appendix C, Eq. (53)] Theorem 1 states that the augmented Lagrangian sequence converges for any sufficiently large ρ, but the proof's condition (53) depends on the trajectory itself: the numerator contains successive differences of the dual variables Φ, μ, z, and the denominator contains differences of a ⊙ ŵ_{n−1}. All of these depend on ρ and on the iterations, and the proof does not establish that a finite ρ satisfying (53) exists or that the ratio in (53) is bounded independently of ρ. As written, the convergence result is circular and does not provide a verifiable choice of ρ. A constructive bound or a different convergence argument is needed.
  3. [Section IV-A, Tables II and III] The synthetic results are reported for a single run with no variation across random seeds or data realizations. The claimed F-score improvements (e.g., 0.60 vs. 0.40 at SR=1; 0.59 vs. 0.39 at σn=0.1) would be more convincing with multiple trials and mean±std or confidence intervals. This is particularly important because the proposed method uses a hand-set degree constraint d=1 and fixed ν=3, and the sensitivity of the F-score to these choices is not reported.
  4. [Section IV-B, portfolio backtest] The MTVGRP portfolio design appears in-sample: the graph L_n is learned from the same data frame on which the portfolio return and Sharpe ratio are then computed, and no train/test split or walk-forward protocol is described. The reported Sharpe ratio of 1.31 (vs. 1.00 for MSRP and 1.25 for EWP) may reflect look-ahead bias and is not evidence of out-of-sample performance. Please clarify whether the data used to estimate the graph and covariance are disjoint from the evaluation period, or provide a proper rolling-window backtest.
minor comments (5)
  1. [Section III, Eq. (21)] In the definition of c_l, the term α^l is used before it is defined (α^l appears in a_l and b_l, but α is introduced in the update (28)). Please reorder or rename to avoid confusion.
  2. [Section III, Eq. (20)] The update (20) uses a hard-threshold 1(c_l > c_th) ⊙ c_l, but the surrounding text refers to l0 regularization. The threshold value c_th = sqrt(2β/(ρ(4p−1))) should be derived explicitly from the proximal operator; please add one line to justify the factor 2.
  3. [Section IV-B, first paragraph] The list of sectors contains "Real State" — this appears to be a typo for "Real Estate." Also, the sentence "The ground-truth labels of the sectors are determined by the GICS classification standard" should include a citation or URL for GICS as in other places.
  4. [Section IV-B, Table IV] The comparison in Table IV uses static benchmark methods with a warm start w0_{n+1} = ŵ_n, while the proposed method uses the temporal VAR model. Please state explicitly how the warm start interacts with the benchmarks' objective, since it may give them information they would not have in a purely static setting.
  5. [General] Several equations use x_t in `Τν` after the update, e.g., Eq. (21) defines Τν but uses x_t in the numerator; check that all quantities are evaluated at the correct iteration index l or l+1.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the MAP derivation from Eqs. (8)-(11) to Eq. (13) is self-contained, and the in-sample evaluation choices do not reduce the central claim to its inputs.

full rationale

The paper's derivation chain is not circular. The negative logarithm of the posterior built from the Student-t likelihood in Eq. (10), the Laplace innovation prior in Eq. (9), the exponential prior on a, and the Gaussian noise model in Eq. (11) is expanded in Appendix A and leads directly to the MAP objective in Eq. (13). Algorithm 1 then minimizes this objective by ADMM/MM updates; nothing in the optimization is defined in terms of the clustering labels or the final performance metrics. The clustering evaluation uses the k-component/rank constraint and spectral clustering, but the graph weights themselves are learned without access to the GICS labels, so the reported ACC/ARI values are not forced by construction. The choices flagged by the reader (fitting nu on the same S&P 500 data, setting k from GICS ground truth, and learning a on the frame being evaluated) are in-sample statistical choices that can weaken external validity, but they are not instances of a fitted parameter being renamed as a prediction or of an output being equivalent to an input. The paper's self-citations, including the conference version [1] and the benchmark [50], are not load-bearing for the theoretical derivation. The most serious technical concern is a correctness/misspecification issue rather than a circularity: for a rank p-k Laplacian, the density in Eq. (10) has tail exponent (nu+p)/2 while the synthetic generator x_t = L^dagger^{1/2} nu_t with nu_t ~ St(0,I) produces a distribution with exponent (nu+p-k)/2 on the range of L; this means the synthetic validation does not sample from the stated density, but Eq. (13) still follows from Eq. (10) as written, so there is no circular reduction to exhibit. Overall, the paper presents an independent derivation supported by numerical experiments, with no circular step that meets the required evidentiary standard.

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

The paper introduces modeling constructs and the MTVGRP portfolio objective, but no new physical entities, mediators, forces, or conserved quantities. The free parameters listed are the quantities the central result depends on that are either fit to data or chosen by hand; the axioms are the unproved modeling and technical premises behind the MAP formulation and the convergence claim.

free parameters (6)
  • nu (Student-t degrees of freedom) = 3 (synthetic); fitted via fitHeavyTail R-package for real data
    Controls the heavy-tailedness of the signal likelihood in Eq. (10). In Section IV.B it is fit to the same S&P 500 data on which clustering is evaluated.
  • a (non-negative VAR coefficient vector) = Estimated per frame, not reported
    Drives edge-weight evolution via w_n = (a elementwise w_{n-1} + epsilon_n)_+ in Eq. (8). It is learned from the current frame by MAP, so it is a per-frame fitted parameter of large dimension (p(p-1)/2).
  • sigma_epsilon (Laplace innovation scale) = exp(0.005 T_n)
    Set by hand in Section IV.A. It determines alpha = 2/(T_n sigma_epsilon) and beta = 2 log(sigma_epsilon)/T_n, which control temporal and graph sparsity.
  • gamma (exponential prior rate on a, scaled) = 0.01
    Regularization on VAR coefficients, set by hand in Section IV.A. It is equivalent to lambda = gamma T_n/2 in the original exponential prior.
  • rho (ADMM penalty parameter) = 3
    ADMM step-size, set by hand in Section IV.A. Theorem 1 only guarantees convergence for rho satisfying a non-constructive condition in Eq. (53).
  • d (node degree vector) = 1 (all entries)
    Equality constraint d_w = d in the feasible set Omega_w. Chosen by hand, not estimated; it strongly restricts the feasible graph and shapes the clustering result.
assumptions (6)
  • domain assumption Signal likelihood is the multivariate Student-t with Laplacian precision: p(x_t|w_n) proportional to det*(L w_n)^{1/2}(1 + x_t^T L w_n x_t / nu)^{-(nu+p)/2} (Eq. 10).
    The entire MAP objective in Proposition 1 rests on this generative form. If financial returns do not follow this heavy-tailed structure, the estimator is misspecified.
  • domain assumption Edge weights evolve as a non-negative VAR: w_n = (a elementwise w_{n-1} + epsilon_n)_+ with Laplace epsilon_n and an exponential prior on a (Eqs. 8-9).
    The temporal consistency term ||w_n - a elementwise w_hat_{n-1}||_1 assumes this specific evolution mechanism, with no evidence that real graph weights obey it.
  • domain assumption Measurement model y_t = m_t elementwise (x_t + n_t) with i.i.d. Gaussian noise n_t of known variance (Eq. 11).
    Used to derive the quadratic fidelity term. Experiments treat sigma_n^2 as known for both synthetic and real data.
  • ad hoc to paper The graph is piecewise constant within frames and the feasible set is Omega_w = {w | rank(L_w) = p-k, d_w = d}.
    The rank constraint enables k-component clustering but is imposed rather than learned. It is the mechanism behind the clustering claims and is not derived from data.
  • standard math Majorization inequalities: log(z) <= log(z0)+(z-z0)/z0 and (w-w^l)^T(zeta I - H)(w-w^l) >= 0 for zeta >= lambda_max(H) (Eqs. 18-19).
    Standard convex majorization tools used to construct MM surrogates in the w and x updates.
  • ad hoc to paper ADMM convergence for a nonconvex objective with an l0 penalty and a discontinuity requires a sufficiently large rho satisfying condition (53) in Appendix C.
    The proof invokes a non-constructive condition that is not shown to hold for any fixed finite rho, so the convergence theorem is not established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Time-Varying Graph Learning for Data with Heavy-Tailed Distribution." pith.science (2026). https://pith.science/paper/UC3IF5M6

@misc{pith2026250100606,
  author       = {Pith},
  title        = {Pith review of: Time-Varying Graph Learning for Data with Heavy-Tailed Distribution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UC3IF5M6}},
  note         = {Machine review of arXiv:2501.00606}
}
read the original abstract

Graph models provide efficient tools to capture the underlying structure of data defined over networks. Many real-world network topologies are subject to change over time. Learning to model the dynamic interactions between entities in such networks is known as time-varying graph learning. Current methodology for learning such models often lacks robustness to outliers in the data and fails to handle heavy-tailed distributions, a common feature in many real-world datasets (e.g., financial data). This paper addresses the problem of learning time-varying graph models capable of efficiently representing heavy-tailed data. Unlike traditional approaches, we incorporate graph structures with specific spectral properties to enhance data clustering in our model. Our proposed method, which can also deal with noise and missing values in the data, is based on a stochastic approach, where a non-negative vector auto-regressive (VAR) model captures the variations in the graph and a Student-t distribution models the signal originating from this underlying time-varying graph. We propose an iterative method to learn time-varying graph topologies within a semi-online framework where only a mini-batch of data is used to update the graph. Simulations with both synthetic and real datasets demonstrate the efficacy of our model in analyzing heavy-tailed data, particularly those found in financial markets.

Figures

Figures reproduced from arXiv: 2501.00606 by the authors.

Figure 2
Figure 2. Illustration of the time frames. vector of all possible edge weights wn ∈ R p(p−1)/2 + . The edge weights can also be mapped to the adjacency matrix using the adjacency operator A [18], i.e., Wn = Awn. Furthermore, the time-varying Laplacian matrix can be derived as Ln = Diag(Wn1) − Wn = Lwn, where L denotes the Laplacian operator [18]. Considering T snapshots (time measurements) of the signal vertically arranged in… view at source ↗
Figure 1
Figure 1. Illustration of the concept of time-varying graphs. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Visualization of the learned (weighted) adjacency matrices at different time intervals, with frames of length [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The k-component graphs learned from financial data corresponding to the log-returns of 100 stocks in the S&P 500 index (comprising k = 8 sectors). The graphs are shown for the last data frame (with length Tn = 200). B. Real Data In this section, we utilize real-world d…
Figure 5
Figure 5. Figure 5: The time evolution of the graphs learned from S&P500 data via the proposed method for the frame length of [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Performance of the proposed method for clustering SP500 data versus [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Sharpe ratio (left) and the maximum draw-down (right) performance [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 58 canonical work pages

  1. [1]

    Learning time-varying graphs for heavy-tailed data clustering,

    Amirhossein Javaheri and Daniel P. Palomar, “Learning time-varying graphs for heavy-tailed data clustering,” in 2024 32nd European Signal Processing Conference (EUSIPCO), 2024, pp. 2472–2476

  2. [2]

    Social Network Analysis with Content and Graphs,

    W. Campbell, C. Dagli, and C. Weinstein, “Social Network Analysis with Content and Graphs,” Lincoln Laboratory Journal , vol. 20, no. 1, pp. 62–81, 2013

  3. [3]

    Improved clustering algorithms for image segmentation based on non-local in- formation and back projection,

    X. Zhang, Y . Sun, H. Liu, Z. Hou, F. Zhao, and C. Zhang, “Improved clustering algorithms for image segmentation based on non-local in- formation and back projection,” Information Sciences , vol. 550, pp. 129–144, Mar. 2021

  4. [4]

    On Applications of Graph/Network Theory to Problems in Communication Systems,

    H. Tamura, K. Nakano, M. Sengoku, and S. Shinoda, “On Applications of Graph/Network Theory to Problems in Communication Systems,” ECTI Transactions on Computer and Information Technology (ECTI- CIT), vol. 5, no. 1, pp. 15–21, Jan. 1970

  5. [5]

    Learning undirected graphs in financial markets,

    J. V . de M. Cardoso and D. P. Palomar, “Learning undirected graphs in financial markets,” in 2020 54th Asilomar Conference on Signals, Systems, and Computers , 2020, pp. 741–745

  6. [6]

    Graph Signal Processing: Overview, Challenges, and Applications,

    A. Ortega, P. Frossard, J. Kova ˇcevi´c, J. M. F. Moura, and P. Van- dergheynst, “Graph Signal Processing: Overview, Challenges, and Applications,” Proceedings of the IEEE , vol. 106, no. 5, pp. 808–828, May 2018

  7. [7]

    Learning graphs from data: A signal representation perspective,

    X. Dong, D. Thanou, M. Rabbat, and P. Frossard, “Learning graphs from data: A signal representation perspective,” IEEE Signal Processing Magazine, vol. 36, no. 3, pp. 44–63, May 2019

  8. [8]

    Rue and L

    H. Rue and L. Held, Gaussian Markov random fields: theory and applications, Number 104 in Monographs on statistics and applied probability. Chapman & Hall/CRC, Boca Raton, 2005

Show all 58 references
  1. [9]

    Nonconvex sparse graph learning under Laplacian constrained graphical model,

    J. Ying, J. V . de M. Cardoso, and D. P. Palomar, “Nonconvex sparse graph learning under Laplacian constrained graphical model,” in Advances in Neural Information Processing Systems , 2020, vol. 33, pp. 7101–7113

  2. [10]

    Minimax estimation of Laplacian constrained precision matrices,

    J. Ying, J. V . de M. Cardoso, and D. P. Palomar, “Minimax estimation of Laplacian constrained precision matrices,” in International Conference on Artificial Intelligence and Statistics , 2021, vol. 130, pp. 3736–3744

  3. [11]

    Graph Learning From Data Under Laplacian and Structural Constraints,

    H. E. Egilmez, E. Pavez, and A. Ortega, “Graph Learning From Data Under Laplacian and Structural Constraints,” IEEE Journal of Selected Topics in Signal Processing , vol. 11, no. 6, pp. 825–841, Sept. 2017

  4. [12]

    Graph signal processing - a probabilistic framework,

    C. Zhang, D. Florencio, and P. A. Chou, “Graph signal processing - a probabilistic framework,” Tech. Rep. MSR-TR-2015-31, April 2015

  5. [13]

    Does the ℓ1-norm learn a sparse graph under Laplacian constrained graphical models?,

    J. Ying, J. V . de M. Cardoso, and D. P. Palomar, “Does the ℓ1-norm learn a sparse graph under Laplacian constrained graphical models?,” arXiv preprint arXiv:2006.14925 , 2020

  6. [14]

    Sparse inverse covariance estimation with the graphical lasso,

    J. Friedman, T. Hastie, and R. Tibshirani, “Sparse inverse covariance estimation with the graphical lasso,” Biostatistics, vol. 9, no. 3, pp. 432–441, July 2008

  7. [15]

    Discovering structure by learning sparse graphs,

    B. Lake and J. Tenenbaum, “Discovering structure by learning sparse graphs,” in Proceedings of the 32nd Annual Meeting of the Cognitive Science Society, Portland, Oregon, United States, Aug. 2010, pp. 778– 784

  8. [16]

    Optimization Algorithms for Graph Laplacian Estimation via ADMM and MM,

    L. Zhao, Y . Wang, S. Kumar, and D. P. Palomar, “Optimization Algorithms for Graph Laplacian Estimation via ADMM and MM,” IEEE Transactions on Signal Processing, vol. 67, no. 16, pp. 4231–4244, Aug. 2019

  9. [17]

    Learning bipartite graphs: Heavy tails and multiple components,

    J. V . de M. Cardoso, J. Ying, and D. P. Palomar, “Learning bipartite graphs: Heavy tails and multiple components,” in Advances in Neural Information Processing Systems , 2022, vol. 35, pp. 14044–14057

  10. [18]

    A Unified Framework for Structured Graph Learning via Spectral Constraints,

    S. Kumar, J. Ying, J. V . de M. Cardoso, and D. P. Palomar, “A Unified Framework for Structured Graph Learning via Spectral Constraints,” Journal of Machine Learning Research, vol. 21, no. 22, pp. 1–60, 2020

  11. [19]

    Kaplan, Structural Equation Modeling (2nd ed.): Foundations and Extensions, SAGE Publications, Inc., 2455 Teller Road, Thousand Oaks California 91320 United States, 2009

    D. Kaplan, Structural Equation Modeling (2nd ed.): Foundations and Extensions, SAGE Publications, Inc., 2455 Teller Road, Thousand Oaks California 91320 United States, 2009

  12. [20]

    Topology Selection in Graphical Models of Autoregressive Processes,

    J. Songsiri and L. Vandenberghe, “Topology Selection in Graphical Models of Autoregressive Processes,” Journal of Machine Learning Research, vol. 11, no. 91, pp. 2671–2705, 2010

  13. [21]

    Causal Network Inference Via Group Sparse Regularization,

    A. Bolstad, B. D. Van Veen, and R. Nowak, “Causal Network Inference Via Group Sparse Regularization,” IEEE Transactions on Signal Processing, vol. 59, no. 6, pp. 2628–2641, June 2011

  14. [22]

    Signal Processing on Graphs: Causal Mod- eling of Unstructured Data,

    J. Mei and J. M. F. Moura, “Signal Processing on Graphs: Causal Mod- eling of Unstructured Data,” IEEE Transactions on Signal Processing , vol. 65, no. 8, pp. 2077–2092, Apr. 2017

  15. [23]

    Learning Spatiotemporal Graphical Models From Incomplete Observations,

    A. Javaheri, A. Amini, F. Marvasti, and D. P. Palomar, “Learning Spatiotemporal Graphical Models From Incomplete Observations,”IEEE Transactions on Signal Processing , vol. 72, pp. 1361–1374, 2024. 14

  16. [24]

    Learning time varying graphs,

    V . Kalofolias, A. Loukas, D. Thanou, and P. Frossard, “Learning time varying graphs,” in 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , New Orleans, LA, Mar. 2017, pp. 2826–2830

  17. [25]

    Network inference via the time-varying graphical lasso,

    D. Hallac, Y . Park, S. Boyd, and J. Leskovec, “Network inference via the time-varying graphical lasso,” in Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY , USA, 2017, p. 205–213

  18. [26]

    Time-Varying Graph Learning with Constraints on Graph Temporal Variation,

    K. Yamada, Y . Tanaka, and A. Ortega, “Time-Varying Graph Learning with Constraints on Graph Temporal Variation,” Jan. 2020, arXiv:2001.03346

  19. [27]

    Learning Undirected Graphs in Financial Markets,

    J. V . de M. Cardoso and D. P. Palomar, “Learning Undirected Graphs in Financial Markets,” in 2020 54th Asilomar Conference on Signals, Systems, and Computers , Nov. 2020, pp. 741–745

  20. [28]

    Online Graph Learning From Time-Varying Structural Equation Models,

    A. Natali, E. Isufi, M. Coutino, and G. Leus, “Online Graph Learning From Time-Varying Structural Equation Models,” in2021 55th Asilomar Conference on Signals, Systems, and Computers , Oct. 2021, pp. 1579– 1585, ISSN: 2576-2303

  21. [29]

    Tracking of a dynamic graph using a signal theory approach : application to the study of a bike sharing system,

    R. Hamon, P. Borgnat, P. Flandrin, and C. Robardet, “Tracking of a dynamic graph using a signal theory approach : application to the study of a bike sharing system,” Sept. 2013, p. 101

  22. [30]

    Dynamic graph models,

    F. Harary and G. Gupta, “Dynamic graph models,” Mathematical and Computer Modelling, vol. 25, no. 7, pp. 79–87, Apr. 1997

  23. [31]

    Time-varying Graph Learning Based on Sparseness of Temporal Variation,

    K. Yamada, Y . Tanaka, and A. Ortega, “Time-varying Graph Learning Based on Sparseness of Temporal Variation,” in ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 2019, pp. 5411–5415

  24. [32]

    Learning Laplacian Matrix in Smooth Graph Signal Representations,

    X. Dong, D. Thanou, P. Frossard, and P. Vandergheynst, “Learning Laplacian Matrix in Smooth Graph Signal Representations,” IEEE Transactions on Signal Processing, vol. 64, no. 23, pp. 6160–6173, Dec. 2016

  25. [33]

    Learning Time-Varying Graphs From Online Data,

    A. Natali, E. Isufi, M. Coutino, and G. Leus, “Learning Time-Varying Graphs From Online Data,” IEEE Open Journal of Signal Processing , vol. 3, pp. 212–228, 2022

  26. [34]

    A Class of Prediction-Correction Methods for Time-Varying Convex Optimization,

    A. Simonetto, A. Mokhtari, A. Koppel, G. Leus, and A. Ribeiro, “A Class of Prediction-Correction Methods for Time-Varying Convex Optimization,” IEEE Transactions on Signal Processing , vol. 64, no. 17, pp. 4576–4591, Sept. 2016

  27. [35]

    Tracking Switched Dynamic Network Topologies From Information Cascades,

    B. Baingana and G. B. Giannakis, “Tracking Switched Dynamic Network Topologies From Information Cascades,” IEEE Transactions on Signal Processing , vol. 65, no. 4, pp. 985–997, Feb. 2017

  28. [36]

    Online Graph Learning from Sequential Data,

    S. Vlaski, H. P. Mareti ´c, R. Nassif, P. Frossard, and A. H. Sayed, “Online Graph Learning from Sequential Data,” in 2018 IEEE Data Science Workshop (DSW), June 2018, pp. 190–194

  29. [37]

    Online Topology Inference from Stream- ing Stationary Graph Signals with Partial Connectivity Information,

    R. Shafipour and G. Mateos, “Online Topology Inference from Stream- ing Stationary Graph Signals with Partial Connectivity Information,” Algorithms, vol. 13, no. 9, pp. 228, Sept. 2020

  30. [38]

    Online Non-linear Topology Identification from Graph-connected Time Series,

    R. Money, J. Krishnan, and B. Beferull-Lozano, “Online Non-linear Topology Identification from Graph-connected Time Series,” in 2021 IEEE Data Science and Learning Workshop (DSLW) , June 2021, pp. 1–6

  31. [39]

    Online Graph Learning under Smoothness Priors,

    S. S. Saboksayr, G. Mateos, and M. Cetin, “Online Graph Learning under Smoothness Priors,” 2021 29th European Signal Processing Conference (EUSIPCO), pp. 1820–1824, Aug. 2021

  32. [40]

    Online learning of time- varying signals and graphs,

    S. Sardellitti, S. Barbarossa, and P. Di Lorenzo, “Online learning of time- varying signals and graphs,” in ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 5230–5234

  33. [41]

    Online Network Inference from Graph-Stationary Signals with Hidden Nodes,

    A. Buciulea, M. Navarro, S. Rey, S. Segarra, and A. G. Marques, “Online Network Inference from Graph-Stationary Signals with Hidden Nodes,” Sept. 2024, arXiv:2409.08760

  34. [42]

    S. I. Resnick, Heavy-tail phenomena: probabilistic and statistical mod- eling, Springer series in operations research and financial engineering. Springer, New York, NY [Heidelberg], 2007

  35. [43]

    Graph-based Methods for Visualization and Clustering,

    J. Paratte, “Graph-based Methods for Visualization and Clustering,” 2017, Publisher: Lausanne, EPFL

  36. [44]

    Joint Signal Recovery and Graph Learning from Incomplete Time-Series,

    A. Javaheri, A. Amini, F. Marvasti, and D. P. Palomar, “Joint Signal Recovery and Graph Learning from Incomplete Time-Series,” inICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , Apr. 2024, pp. 13511–13515

  37. [45]

    Graphical Models in Heavy-Tailed Markets,

    J. V . de M. Cardoso, J. Ying, and D. P. Palomar, “Graphical Models in Heavy-Tailed Markets,” in Advances in Neural Information Processing Systems, 2021, vol. 34, pp. 19989–20001

  38. [46]

    Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers,

    S. Boyd, “Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers,” Foundations and Trends® in Machine Learning , vol. 3, no. 1, pp. 1–122, 2010

  39. [47]

    Majorization-Minimization Algo- rithms in Signal Processing, Communications, and Machine Learning,

    Y . Sun, P. Babu, and D. P. Palomar, “Majorization-Minimization Algo- rithms in Signal Processing, Communications, and Machine Learning,” IEEE Transactions on Signal Processing , vol. 65, no. 3, pp. 794–816, Feb. 2017

  40. [48]

    Regression Shrinkage and Selection Via the Lasso,

    R. Tibshirani, “Regression Shrinkage and Selection Via the Lasso,” Journal of the Royal Statistical Society: Series B (Methodological) , vol. 58, no. 1, pp. 267–288, 1996

  41. [49]

    The Constrained Laplacian Rank algorithm for graph-based clustering,

    F. Nie, X. Wang, M. I. Jordan, and H. Huang, “The Constrained Laplacian Rank algorithm for graph-based clustering,” in Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence , Phoenix, Arizona, Feb. 2016, AAAI’16, pp. 1969–1976

  42. [50]

    Graph Learning for Balanced Clustering of Heavy-Tailed Data,

    A. Javaheri, J. V . de M. Cardoso, and D. P. Palomar, “Graph Learning for Balanced Clustering of Heavy-Tailed Data,” in 2023 IEEE 9th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), Herradura, Costa Rica, Dec. 2023, pp. 481–485

  43. [51]

    Everitt, Ed., Cluster analysis , Wiley series in probability and statistics

    B. Everitt, Ed., Cluster analysis , Wiley series in probability and statistics. Wiley, Chichester, West Sussex, U.K, 5th ed edition, 2011

  44. [52]

    Modularity and community structure in networks,

    M. E. J. Newman, “Modularity and community structure in networks,” Proceedings of the National Academy of Sciences , vol. 103, no. 23, pp. 8577–8582, June 2006

  45. [53]

    Objective Criteria for the Evaluation of Clustering Methods,

    W. M. Rand, “Objective Criteria for the Evaluation of Clustering Methods,” Journal of the American Statistical Association , vol. 66, no. 336, pp. 846–850, Dec. 1971

  46. [54]

    On Spectral Clustering: Analysis and an algorithm,

    A. Ng, M. Jordan, and Y . Weiss, “On Spectral Clustering: Analysis and an algorithm,” in Advances in Neural Information Processing Systems , 2001, NIPS’01, p. 849–856

  47. [55]

    Some methods for classification and analysis of multi- variate observations,

    J. MacQueen, “Some methods for classification and analysis of multi- variate observations,” in Proc. 5th Berkeley Symp. Math. Stat. Probab. , 1967, pp. 481–485

  48. [56]

    Mutual Fund Performance,

    W. F. Sharpe, “Mutual Fund Performance,” The Journal of Business , vol. 39, no. 1, pp. 119–138, 1966

  49. [57]

    Parameter-free convex equivalent and dual programs of fractional programming problems,

    S. Schaible, “Parameter-free convex equivalent and dual programs of fractional programming problems,” Zeitschrift f¨ur Operations Research, vol. 18, no. 5, pp. 187–196, Oct. 1974

  50. [58]

    A. J. Laub, Matrix analysis for scientists and engineers , Society for Industrial and Applied Mathematics, Philadelphia, 2005. 15

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.