Copula-Induced Correntropy for Robust Conjugate Gradient Learning

David Morales-Jimenez; Farshad Rostami Ghadi; F. Javier Lopez-Martinez; Kai-Kit Wong; Marios Kountouris

arxiv: 2605.23044 · v1 · pith:IV5KET56new · submitted 2026-05-21 · 📡 eess.SP

Copula-Induced Correntropy for Robust Conjugate Gradient Learning

Farshad Rostami Ghadi , F. Javier Lopez-Martinez , David Morales-Jimenez , Kai-Kit Wong , Marios Kountouris This is my paper

Pith reviewed 2026-05-25 05:13 UTC · model grok-4.3

classification 📡 eess.SP

keywords copula-induced correntropyinformation-theoretic learningrobust conjugate gradientdependent heavy-tailed noisemultivariate signal regressionadaptive filtering

0 comments

The pith

Defining correntropy in copula-transformed residual space separates marginal robustness from dependence weighting for improved conjugate gradient learning.

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

The paper establishes a copula-induced information-theoretic learning framework that extends correntropy to explicitly capture dependence among residuals in multivariate signals. By transforming residuals into copula space, the criterion handles marginal distributions separately from their joint dependence structure, unlike componentwise correntropy. A mixed marginal-dependence objective is derived along with a tailored conjugate gradient algorithm that carries convergence guarantees under fixed estimators. Experiments on synthetic regression tasks show the approach outperforms MSE, Huber, Student's-t, and classical correntropy, especially under dependent heavy-tailed noise. The work targets robust adaptive learning in signal processing where noise exhibits statistical dependence.

Core claim

The central claim is that a copula-induced correntropy objective, defined on copula-transformed residuals rather than raw residuals, produces a learning criterion that separates marginal robustness from dependence weighting and yields a robust conjugate gradient algorithm with sufficient descent and stationarity guarantees for fixed smooth marginal estimators and fixed copula metrics.

What carries the argument

Copula-induced correntropy (CIC) objective, which embeds a copula space representation of residual dependence into the similarity measure while using a mixed marginal-dependence objective.

If this is right

The method supplies information-theoretic and Bayesian interpretations of the new criterion.
Sufficient descent and global stationarity are guaranteed for the fixed-estimator subproblem under standard line-search conditions.
A robust conjugate gradient algorithm is obtained that is tailored to the copula-induced criterion.
Consistent outperformance holds over MSE, Huber, Student's-t, and classical correntropy in synthetic multivariate regression with dependent heavy-tailed noise.

Where Pith is reading between the lines

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

The modular separation of marginal and dependence terms may permit independent tuning of each component in other adaptive filtering settings.
Testing on real multi-sensor time-series data would reveal whether the synthetic gains translate when dependence structures are unknown and time-varying.
Relaxing the fixed-estimator assumption to allow online marginal adaptation could extend applicability to non-stationary environments.

Load-bearing premise

A fixed smooth marginal estimator, a fixed copula-space metric, and a regularized radial penalty are sufficient to separate marginal robustness from dependence weighting in a way that improves learning.

What would settle it

A controlled synthetic multivariate regression experiment with dependent heavy-tailed noise in which the proposed method shows no performance gain over classical correntropy or MSE would falsify the central performance claim.

Figures

Figures reproduced from arXiv: 2605.23044 by David Morales-Jimenez, Farshad Rostami Ghadi, F. Javier Lopez-Martinez, Kai-Kit Wong, Marios Kountouris.

**Figure 5.** Figure 5: Ablation study on tail robustness under dependent he [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗

**Figure 4.** Figure 4: Distribution of absolute prediction errors [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗

read the original abstract

Robust learning in the presence of non-Gaussian and statistically dependent noise remains a fundamental challenge in signal processing and adaptive systems. Although information-theoretic learning criteria such as correntropy offer strong robustness against impulsive and heavy-tailed disturbances, existing formulations are commonly applied componentwise and therefore do not explicitly exploit the dependence structures inherent in multivariate, multi-sensor, and temporal signals. In this paper, we propose a learning framework, termed \textit{copula-induced information-theoretic learning} (CITL), which extends correntropy by embedding a copula space representation of residual dependence into the similarity measure. Unlike conventional correntropy-based approaches that operate pointwise on raw residuals, the proposed criterion is defined in a copula-transformed residual space, thus separating marginal robustness from dependence weighting. We derive a copula-induced correntropy (CIC) objective and a mixed marginal--dependence objective used in the implementation, provide information-theoretic and Bayesian interpretations, and develop a robust conjugate gradient (CG) learning algorithm tailored to this criterion. For fixed smooth marginal estimators, a fixed copula-space metric, and a regularized radial penalty, we establish sufficient descent and global stationarity guarantees for the corresponding fixed-estimator subproblem under standard line-search conditions. Experiments on synthetic multivariate signal processing regression problems demonstrate that the proposed method consistently outperforms mean squared error (MSE), Huber, Student's-$t$, and classical correntropy-based approaches, particularly in the presence of dependent heavy-tailed noise.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper adds copula dependence modeling to correntropy for multivariate signal regression, but the stationarity proof covers only the fixed-marginal subproblem while experiments run the full adaptive CG.

read the letter

The core new piece is the CITL criterion that maps residuals through copulas before applying a correntropy-style similarity, so the method can weight dependence separately from marginal tails. That is a clean way to extend existing correntropy work to cases where noise across sensors or time steps is correlated and heavy-tailed. They also supply information-theoretic and Bayesian readings of the objective and a conjugate-gradient solver with a regularized radial term. For the restricted case of fixed smooth marginal estimators and fixed copula metric, they show sufficient descent and stationarity under line search; that part looks formally grounded on its own terms. The experiments claim consistent gains over MSE, Huber, t-distribution, and plain correntropy on synthetic multivariate regression with dependent impulsive noise, which matches the motivation. The main gap is the one flagged in the stress test: the stationarity result does not automatically apply once the algorithm starts adapting or jointly estimating the marginals and copula parameters, yet that is what the reported CG procedure must do to exploit the dependence structure. The abstract gives no numbers, error bars, or details on copula selection, so the empirical strength cannot be judged from the summary alone. If the full paper shows that the adaptive steps still converge in practice or provides a separate argument, the concern shrinks; otherwise the link between theory and results stays loose. This is aimed at the adaptive signal-processing community that already uses correntropy or M-estimators. A reader working on robust multi-sensor or temporal filtering could extract the construction and test it on their own data. The work is coherent enough on its own terms to deserve referee time, even if revisions will be needed on the theory-experiment connection.

Referee Report

1 major / 1 minor

Summary. The manuscript proposes copula-induced information-theoretic learning (CITL) as an extension of correntropy that embeds a copula-space representation of residual dependence to handle dependent heavy-tailed noise in multivariate signal processing tasks. It derives a copula-induced correntropy (CIC) objective together with a mixed marginal-dependence formulation, supplies information-theoretic and Bayesian interpretations, develops a robust conjugate gradient algorithm, and establishes sufficient descent plus global stationarity for the fixed-estimator subproblem (under fixed smooth marginal estimators, fixed copula metric, and regularized radial penalty). Synthetic regression experiments are reported to show consistent outperformance versus MSE, Huber, Student's-t, and classical correntropy baselines, especially under dependent heavy-tailed noise.

Significance. If the empirical superiority is confirmed with quantitative detail and the stationarity result can be connected to the adaptive algorithm actually used, the separation of marginal robustness from dependence weighting would constitute a useful conceptual advance for robust adaptive filtering and multi-sensor processing. The provision of stationarity guarantees, even if limited to the fixed-estimator subproblem, is a positive technical feature that distinguishes the work from purely heuristic robust criteria.

major comments (1)

[Abstract] Abstract (and the section deriving the stationarity result): sufficient descent and global stationarity are established only for the fixed-estimator subproblem with fixed smooth marginal estimators, fixed copula-space metric, and regularized radial penalty. The central experimental claim concerns performance of the full robust conjugate gradient learning algorithm on synthetic tasks with dependent heavy-tailed noise. If the implemented procedure adapts or jointly estimates the marginals and copula parameters (as would be necessary to exploit residual dependence), the proven properties do not automatically transfer, leaving the link between the theoretical guarantees and the reported outperformance unestablished.

minor comments (1)

[Abstract] The abstract states that the method 'consistently outperforms' the baselines but supplies no quantitative metrics, error bars, or description of how the copula family and marginal estimators were selected; these details should be added to the experimental section for reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback and positive assessment of the conceptual contribution. We address the major comment below.

read point-by-point responses

Referee: [Abstract] Abstract (and the section deriving the stationarity result): sufficient descent and global stationarity are established only for the fixed-estimator subproblem with fixed smooth marginal estimators, fixed copula-space metric, and regularized radial penalty. The central experimental claim concerns performance of the full robust conjugate gradient learning algorithm on synthetic tasks with dependent heavy-tailed noise. If the implemented procedure adapts or jointly estimates the marginals and copula parameters (as would be necessary to exploit residual dependence), the proven properties do not automatically transfer, leaving the link between the theoretical guarantees and the reported outperformance unestablished.

Authors: We thank the referee for highlighting this distinction. In the robust conjugate gradient algorithm, marginal estimators are computed once via a fixed nonparametric procedure (e.g., kernel density estimation) and held constant thereafter; the copula-space metric is likewise selected and fixed a priori. The CG iterations optimize only the CIC objective under these fixed components, corresponding exactly to the fixed-estimator subproblem for which sufficient descent and global stationarity are established. Dependence is incorporated through the fixed copula without joint adaptation of marginals. We will revise the algorithm description and experimental setup to state this linkage explicitly. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation introduces CITL via copula embedding and separates subproblem analysis from experiments

full rationale

The paper defines a new CITL criterion by embedding a copula-space representation of residual dependence into correntropy, derives the CIC objective and mixed marginal-dependence objective, and states stationarity guarantees explicitly limited to the fixed-estimator subproblem under fixed marginals, fixed metric, and radial penalty. No equations or claims in the provided text reduce the objective to a fitted parameter renamed as prediction, invoke self-citation as load-bearing uniqueness, or smuggle an ansatz. The experimental outperformance claims are presented as separate empirical results on synthetic tasks. The derivation chain remains self-contained against external benchmarks with no self-definitional or construction-equivalent reductions.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are stated beyond the high-level claim that the copula representation separates marginal and dependence effects.

pith-pipeline@v0.9.0 · 5812 in / 1147 out tokens · 20225 ms · 2026-05-25T05:13:08.696354+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages · 1 internal anchor

[1]

R. A. Maronna, R. D. Martin, V . J. Y ohai and M. Salibi´ an-B arrera, Robust Statistics: Theory and Methods (with R) , John Wiley & Sons, 2019

work page 2019
[2]

S. S. Haykin, Adaptive Filter Theory , Pearson Education India, 2002

work page 2002
[3]

Majorization-Minimi zation Algo- rithms in Signal Processing, Communications, and Machine L earning,

Y . Sun, P . Babu, and D. P . Palomar, “Majorization-Minimi zation Algo- rithms in Signal Processing, Communications, and Machine L earning,” IEEE Trans. Signal Process. , vol. 65, no. 3, pp. 794-816, Feb. 2017

work page 2017
[4]

Robust statistics,

P . J. Huber, “Robust statistics,” International encyclopedia of statistical science, Springer, Berlin, Heidelberg, pp. 1248-1251, 2011

work page 2011
[5]

Time -V arying Graph Learning for Data With Heavy-Tailed Distribution,

A. Javaheri, J. Ying, D. P . Palomar and F. Marvasti, “Time -V arying Graph Learning for Data With Heavy-Tailed Distribution,” IEEE Trans. Signal Process. , vol. 73, pp. 3044-3060, 2025

work page 2025
[6]

Impuls ive Noise Modeling and Robust Receiver Design,

L. Clavier, G. W. Peters, F. Septier and I. Nevat, “Impuls ive Noise Modeling and Robust Receiver Design,” EURASIP J. Wirel. Commun. Netw., vol. 13, no. 1, 2021

work page 2021
[7]

Adaptive Lp-norm Diversity Combining in Non-Gaussian Noise and Interference,

A. Nasri, A. Nezampour and R. Schober, “Adaptive Lp-norm Diversity Combining in Non-Gaussian Noise and Interference,” IEEE Trans. Wireless Commun., vol. 8, no. 8, pp. 4230-4240, Aug. 2009

work page 2009
[8]

Huber-based Adap tive Unscented Kalman Filter with Non-Gaussian Measurement Noi se,

B. Zhu, L. Chang, J. Xu, F. Zha and J. Li, “Huber-based Adap tive Unscented Kalman Filter with Non-Gaussian Measurement Noi se,” Circuits Syst. Signal Process. , vol. 37, no. 9, pp. 3842-3861, 2018

work page 2018
[9]

A Generalized t-Dis tribution- Based Kernel Adaptive Filtering Algorithm,

H. Tang, H. Han, S. Zhang and W. Feng, “A Generalized t-Dis tribution- Based Kernel Adaptive Filtering Algorithm,” IEEE Trans. Circuits Syst. II: Express Briefs , vol. 71, no. 6, pp. 3241-3245, June 2024

work page 2024
[10]

A Novel Ro bust Gaussian-Student’s t Mixture Distribution Based Kalman Fi lter,

Y . Huang, Y . Zhang, Y . Zhao and J. A. Chambers, “A Novel Ro bust Gaussian-Student’s t Mixture Distribution Based Kalman Fi lter,” IEEE Trans. Signal Process. , vol. 67, no. 13, pp. 3606-3620, July 2019

work page 2019
[11]

Copulae: An Overview and Rece nt Develop- ments,

J. Gr¨ oßer and O. Okhrin, “Copulae: An Overview and Rece nt Develop- ments,” Wiley Interdiscip. Rev. Comput. Stat. , vol. 14, no. 3, 2022

work page 2022
[12]

J. C. Principe, Information theoretic learning: Renyi’s entropy and kerne l perspectives, Springer Science & Business Media, 2010

work page 2010
[13]

Correntropy: Properties and Applications in Non-Gaussian Signal Processing,

W. Liu, P . P . Pokharel and J. C. Principe, “Correntropy: Properties and Applications in Non-Gaussian Signal Processing,” IEEE Trans. Signal Process., vol. 55, no. 11, pp. 5286-5298, 2007

work page 2007
[14]

A New Correntropy-Bas ed Con- jugate Gradient Backpropagation Algorithm for Improving T raining in Neural Networks,

A. R. Heravi and G. Abed Hodtani, “A New Correntropy-Bas ed Con- jugate Gradient Backpropagation Algorithm for Improving T raining in Neural Networks,” IEEE Trans. Neural Netw. Learn. Syst. , vol. 29, no. 12, pp. 6252-6263, Dec. 2018

work page 2018
[15]

Robustness of Maximum Correntropy Estimation Against Large Outliers

B. Chen, L. Xing, H. Zhao, B. Xu and J. C. Principe, “Robus tness of maximum correntropy estimation against large outliers,” arXiv preprint, https://arxiv.org/abs/1703.08065v2, 2017

work page internal anchor Pith review Pith/arXiv arXiv 2017
[16]

G eneral- ized Correntropy for Robust Adaptive Filtering,

B. Chen, L. Xing, H. Zhao, N. Zheng and J. C. Pr´ ıncipe, “G eneral- ized Correntropy for Robust Adaptive Filtering,” IEEE Trans. Signal Process., vol. 64, no. 13, pp. 3376-3387, July, 2016. 13

work page 2016
[17]

R. B. Nelsen, An Introduction to Copulas , New Y ork, NY: Springer New Y ork, Jan. 2006

work page 2006
[18]

Cop ulas for Statistical Signal Processing (Part I): Extensions and Gen eralization,

X. Zeng, J. Ren, Z. Wang, S. Marshall and T. Durrani, “Cop ulas for Statistical Signal Processing (Part I): Extensions and Gen eralization,” Signal Process. , vol. 94, pp. 691-702, 2014

work page 2014
[19]

Communications meets copula modeling: Non-standard depe ndence features in wireless fading channels,

G. W. Peters, T. A. Myrvoll, T. Matsui, I. Nevat and F. Sep tier, “Communications meets copula modeling: Non-standard depe ndence features in wireless fading channels,” 2014 IEEE Global Conference on Signal and Information Processing (GlobalSIP) , Atlanta, GA, USA, 2014, pp. 1224-1228

work page 2014
[20]

Copula-Based Analy sis of Physical Layer Security Performances Over Correlated Rayl eigh Fading Channels,

F. Rostami Ghadi and G. A. Hodtani, “Copula-Based Analy sis of Physical Layer Security Performances Over Correlated Rayl eigh Fading Channels,” IEEE Transactions on Information F orensics and Security , vol. 16, pp. 431-440, 2021

work page 2021
[21]

Copula- Based Interference Models for IoT Wireless Networks,

C. Zheng, M. Egan, L. Clavier, G. W. Peters and J. -M. Gorc e, “Copula- Based Interference Models for IoT Wireless Networks,” 2019 IEEE International Conference on Communications (ICC) , Shanghai, China, 2019, pp. 1-6

work page 2019
[22]

Copula-Based Bounds for Multi- User Communications–Part I: Average Performance,

E. A. Jorswieck and K. -L. Besser, “Copula-Based Bounds for Multi- User Communications–Part I: Average Performance,” IEEE Communi- cations Letters , vol. 25, no. 1, pp. 3-7, Jan. 2021

work page 2021
[23]

Maximum Correntropy Estima tion Is a Smoothed MAP Estimation,

B. Chen and J. C. Principe, “Maximum Correntropy Estima tion Is a Smoothed MAP Estimation,” IEEE Signal Process. Lett. , vol. 19, no. 8, pp. 491-494, Aug. 2012

work page 2012
[24]

Nocedal and S

J. Nocedal and S. J. Wright, Numerical optimization , New Y ork, NY: Springer New Y ork, Jul. 2006

work page 2006
[25]

A Fletcher-Reeves Conjugate Gradient Neural-Network- Based Localization Algorithm for Wireless Sensor Networks ,

A. Chatterjee, “A Fletcher-Reeves Conjugate Gradient Neural-Network- Based Localization Algorithm for Wireless Sensor Networks ,” IEEE Trans. V eh. Technol., vol. 59, no. 2, pp. 823-830, Feb. 2010

work page 2010
[26]

A Descent Modiﬁed Polak-R ibi´ ere- Polyak Conjugate Gradient Method and Its Global Convergenc e,

L. Zhang, W. Zhou and D. H. Li, “A Descent Modiﬁed Polak-R ibi´ ere- Polyak Conjugate Gradient Method and Its Global Convergenc e,” IMA J. Numer . Anal., vol. 24, no. 6, pp. 629-640, 2006

work page 2006
[27]

A New Conjugate Gradient Metho d with Guaranteed Descent and an Efﬁcient Line Search,

W. W. Hager and H. Zhang, “A New Conjugate Gradient Metho d with Guaranteed Descent and an Efﬁcient Line Search,” SIAM J. Optim. , vol. 16, no. 1, pp. 170-192, 2005

work page 2005
[28]

B. W. Silverman, Density estimation for statistics and data analysis , Routledge, 2018

work page 2018
[29]

A Well-conditioned Estimator fo r Large- Dimensional Covariance Matrices,

O. Ledoit and M. Wolf, “A Well-conditioned Estimator fo r Large- Dimensional Covariance Matrices,” J. Multivar . Anal. , vol. 88, no. 2, pp. 365-411, 2004

work page 2004
[30]

Shrinkage Algorithms for MMSE Covariance Estimation,

Y . Chen, A. Wiesel, Y . C. Eldar and A. O. Hero, “Shrinkage Algorithms for MMSE Covariance Estimation,” IEEE Trans. Signal Process. , vol. 58, no. 10, pp. 5016-5029, Oct. 2010

work page 2010
[31]

Global Convergence Prope rties of Con- jugate Gradient Methods for Optimization,

J. C. Gilbert and J. Nocedal, “Global Convergence Prope rties of Con- jugate Gradient Methods for Optimization,” SIAM J. Optim. , vol. 2, no. 1, pp. 21-42, 1992

work page 1992
[32]

A New Conjugate Gradie nt Method with A Restart Direction and Its Application in Image Restora- tion,

Y . Li, C. Li, W. Y ang and W. Zhang, “A New Conjugate Gradie nt Method with A Restart Direction and Its Application in Image Restora- tion,” AIMS Math , vol. 8, no. 12, pp. 28791-28807, 2023

work page 2023
[33]

Robust estimation of a location parameter ,

P . J. Huber, “Robust estimation of a location parameter ,” Breakthroughs in statistics: Methodology and distribution , pp. 492-518, New Y ork, NY: Springer New Y ork, 1992

work page 1992
[34]

Robust St atistical Modeling Using the t Distribution,

K. L. Lange, R. J. A. Little and J. M. G. Taylor, “Robust St atistical Modeling Using the t Distribution,” J. Am. Stat. Assoc. , vol. 84, no. 408, pp. 881-896, 1989

work page 1989
[35]

W. Liu, J. C. Principe and S. Haykin, Kernel Adaptive Filtering: A Comprehensive Introduction, John Wiley & Sons, 2011

work page 2011
[36]

Identiﬁcation an d control of dynamical systems using neural networks,

K. S. Narendra and K. Parthasarathy, “Identiﬁcation an d control of dynamical systems using neural networks,” IEEE Trans. Neural Netw. , vol. 1, no. 1, pp. 4-27, March 1990

work page 1990

[1] [1]

R. A. Maronna, R. D. Martin, V . J. Y ohai and M. Salibi´ an-B arrera, Robust Statistics: Theory and Methods (with R) , John Wiley & Sons, 2019

work page 2019

[2] [2]

S. S. Haykin, Adaptive Filter Theory , Pearson Education India, 2002

work page 2002

[3] [3]

Majorization-Minimi zation Algo- rithms in Signal Processing, Communications, and Machine L earning,

Y . Sun, P . Babu, and D. P . Palomar, “Majorization-Minimi zation Algo- rithms in Signal Processing, Communications, and Machine L earning,” IEEE Trans. Signal Process. , vol. 65, no. 3, pp. 794-816, Feb. 2017

work page 2017

[4] [4]

Robust statistics,

P . J. Huber, “Robust statistics,” International encyclopedia of statistical science, Springer, Berlin, Heidelberg, pp. 1248-1251, 2011

work page 2011

[5] [5]

Time -V arying Graph Learning for Data With Heavy-Tailed Distribution,

A. Javaheri, J. Ying, D. P . Palomar and F. Marvasti, “Time -V arying Graph Learning for Data With Heavy-Tailed Distribution,” IEEE Trans. Signal Process. , vol. 73, pp. 3044-3060, 2025

work page 2025

[6] [6]

Impuls ive Noise Modeling and Robust Receiver Design,

L. Clavier, G. W. Peters, F. Septier and I. Nevat, “Impuls ive Noise Modeling and Robust Receiver Design,” EURASIP J. Wirel. Commun. Netw., vol. 13, no. 1, 2021

work page 2021

[7] [7]

Adaptive Lp-norm Diversity Combining in Non-Gaussian Noise and Interference,

A. Nasri, A. Nezampour and R. Schober, “Adaptive Lp-norm Diversity Combining in Non-Gaussian Noise and Interference,” IEEE Trans. Wireless Commun., vol. 8, no. 8, pp. 4230-4240, Aug. 2009

work page 2009

[8] [8]

Huber-based Adap tive Unscented Kalman Filter with Non-Gaussian Measurement Noi se,

B. Zhu, L. Chang, J. Xu, F. Zha and J. Li, “Huber-based Adap tive Unscented Kalman Filter with Non-Gaussian Measurement Noi se,” Circuits Syst. Signal Process. , vol. 37, no. 9, pp. 3842-3861, 2018

work page 2018

[9] [9]

A Generalized t-Dis tribution- Based Kernel Adaptive Filtering Algorithm,

H. Tang, H. Han, S. Zhang and W. Feng, “A Generalized t-Dis tribution- Based Kernel Adaptive Filtering Algorithm,” IEEE Trans. Circuits Syst. II: Express Briefs , vol. 71, no. 6, pp. 3241-3245, June 2024

work page 2024

[10] [10]

A Novel Ro bust Gaussian-Student’s t Mixture Distribution Based Kalman Fi lter,

Y . Huang, Y . Zhang, Y . Zhao and J. A. Chambers, “A Novel Ro bust Gaussian-Student’s t Mixture Distribution Based Kalman Fi lter,” IEEE Trans. Signal Process. , vol. 67, no. 13, pp. 3606-3620, July 2019

work page 2019

[11] [11]

Copulae: An Overview and Rece nt Develop- ments,

J. Gr¨ oßer and O. Okhrin, “Copulae: An Overview and Rece nt Develop- ments,” Wiley Interdiscip. Rev. Comput. Stat. , vol. 14, no. 3, 2022

work page 2022

[12] [12]

J. C. Principe, Information theoretic learning: Renyi’s entropy and kerne l perspectives, Springer Science & Business Media, 2010

work page 2010

[13] [13]

Correntropy: Properties and Applications in Non-Gaussian Signal Processing,

W. Liu, P . P . Pokharel and J. C. Principe, “Correntropy: Properties and Applications in Non-Gaussian Signal Processing,” IEEE Trans. Signal Process., vol. 55, no. 11, pp. 5286-5298, 2007

work page 2007

[14] [14]

A New Correntropy-Bas ed Con- jugate Gradient Backpropagation Algorithm for Improving T raining in Neural Networks,

A. R. Heravi and G. Abed Hodtani, “A New Correntropy-Bas ed Con- jugate Gradient Backpropagation Algorithm for Improving T raining in Neural Networks,” IEEE Trans. Neural Netw. Learn. Syst. , vol. 29, no. 12, pp. 6252-6263, Dec. 2018

work page 2018

[15] [15]

Robustness of Maximum Correntropy Estimation Against Large Outliers

B. Chen, L. Xing, H. Zhao, B. Xu and J. C. Principe, “Robus tness of maximum correntropy estimation against large outliers,” arXiv preprint, https://arxiv.org/abs/1703.08065v2, 2017

work page internal anchor Pith review Pith/arXiv arXiv 2017

[16] [16]

G eneral- ized Correntropy for Robust Adaptive Filtering,

B. Chen, L. Xing, H. Zhao, N. Zheng and J. C. Pr´ ıncipe, “G eneral- ized Correntropy for Robust Adaptive Filtering,” IEEE Trans. Signal Process., vol. 64, no. 13, pp. 3376-3387, July, 2016. 13

work page 2016

[17] [17]

R. B. Nelsen, An Introduction to Copulas , New Y ork, NY: Springer New Y ork, Jan. 2006

work page 2006

[18] [18]

Cop ulas for Statistical Signal Processing (Part I): Extensions and Gen eralization,

X. Zeng, J. Ren, Z. Wang, S. Marshall and T. Durrani, “Cop ulas for Statistical Signal Processing (Part I): Extensions and Gen eralization,” Signal Process. , vol. 94, pp. 691-702, 2014

work page 2014

[19] [19]

Communications meets copula modeling: Non-standard depe ndence features in wireless fading channels,

G. W. Peters, T. A. Myrvoll, T. Matsui, I. Nevat and F. Sep tier, “Communications meets copula modeling: Non-standard depe ndence features in wireless fading channels,” 2014 IEEE Global Conference on Signal and Information Processing (GlobalSIP) , Atlanta, GA, USA, 2014, pp. 1224-1228

work page 2014

[20] [20]

Copula-Based Analy sis of Physical Layer Security Performances Over Correlated Rayl eigh Fading Channels,

F. Rostami Ghadi and G. A. Hodtani, “Copula-Based Analy sis of Physical Layer Security Performances Over Correlated Rayl eigh Fading Channels,” IEEE Transactions on Information F orensics and Security , vol. 16, pp. 431-440, 2021

work page 2021

[21] [21]

Copula- Based Interference Models for IoT Wireless Networks,

C. Zheng, M. Egan, L. Clavier, G. W. Peters and J. -M. Gorc e, “Copula- Based Interference Models for IoT Wireless Networks,” 2019 IEEE International Conference on Communications (ICC) , Shanghai, China, 2019, pp. 1-6

work page 2019

[22] [22]

Copula-Based Bounds for Multi- User Communications–Part I: Average Performance,

E. A. Jorswieck and K. -L. Besser, “Copula-Based Bounds for Multi- User Communications–Part I: Average Performance,” IEEE Communi- cations Letters , vol. 25, no. 1, pp. 3-7, Jan. 2021

work page 2021

[23] [23]

Maximum Correntropy Estima tion Is a Smoothed MAP Estimation,

B. Chen and J. C. Principe, “Maximum Correntropy Estima tion Is a Smoothed MAP Estimation,” IEEE Signal Process. Lett. , vol. 19, no. 8, pp. 491-494, Aug. 2012

work page 2012

[24] [24]

Nocedal and S

J. Nocedal and S. J. Wright, Numerical optimization , New Y ork, NY: Springer New Y ork, Jul. 2006

work page 2006

[25] [25]

A Fletcher-Reeves Conjugate Gradient Neural-Network- Based Localization Algorithm for Wireless Sensor Networks ,

A. Chatterjee, “A Fletcher-Reeves Conjugate Gradient Neural-Network- Based Localization Algorithm for Wireless Sensor Networks ,” IEEE Trans. V eh. Technol., vol. 59, no. 2, pp. 823-830, Feb. 2010

work page 2010

[26] [26]

A Descent Modiﬁed Polak-R ibi´ ere- Polyak Conjugate Gradient Method and Its Global Convergenc e,

L. Zhang, W. Zhou and D. H. Li, “A Descent Modiﬁed Polak-R ibi´ ere- Polyak Conjugate Gradient Method and Its Global Convergenc e,” IMA J. Numer . Anal., vol. 24, no. 6, pp. 629-640, 2006

work page 2006

[27] [27]

A New Conjugate Gradient Metho d with Guaranteed Descent and an Efﬁcient Line Search,

W. W. Hager and H. Zhang, “A New Conjugate Gradient Metho d with Guaranteed Descent and an Efﬁcient Line Search,” SIAM J. Optim. , vol. 16, no. 1, pp. 170-192, 2005

work page 2005

[28] [28]

B. W. Silverman, Density estimation for statistics and data analysis , Routledge, 2018

work page 2018

[29] [29]

A Well-conditioned Estimator fo r Large- Dimensional Covariance Matrices,

O. Ledoit and M. Wolf, “A Well-conditioned Estimator fo r Large- Dimensional Covariance Matrices,” J. Multivar . Anal. , vol. 88, no. 2, pp. 365-411, 2004

work page 2004

[30] [30]

Shrinkage Algorithms for MMSE Covariance Estimation,

Y . Chen, A. Wiesel, Y . C. Eldar and A. O. Hero, “Shrinkage Algorithms for MMSE Covariance Estimation,” IEEE Trans. Signal Process. , vol. 58, no. 10, pp. 5016-5029, Oct. 2010

work page 2010

[31] [31]

Global Convergence Prope rties of Con- jugate Gradient Methods for Optimization,

J. C. Gilbert and J. Nocedal, “Global Convergence Prope rties of Con- jugate Gradient Methods for Optimization,” SIAM J. Optim. , vol. 2, no. 1, pp. 21-42, 1992

work page 1992

[32] [32]

A New Conjugate Gradie nt Method with A Restart Direction and Its Application in Image Restora- tion,

Y . Li, C. Li, W. Y ang and W. Zhang, “A New Conjugate Gradie nt Method with A Restart Direction and Its Application in Image Restora- tion,” AIMS Math , vol. 8, no. 12, pp. 28791-28807, 2023

work page 2023

[33] [33]

Robust estimation of a location parameter ,

P . J. Huber, “Robust estimation of a location parameter ,” Breakthroughs in statistics: Methodology and distribution , pp. 492-518, New Y ork, NY: Springer New Y ork, 1992

work page 1992

[34] [34]

Robust St atistical Modeling Using the t Distribution,

K. L. Lange, R. J. A. Little and J. M. G. Taylor, “Robust St atistical Modeling Using the t Distribution,” J. Am. Stat. Assoc. , vol. 84, no. 408, pp. 881-896, 1989

work page 1989

[35] [35]

W. Liu, J. C. Principe and S. Haykin, Kernel Adaptive Filtering: A Comprehensive Introduction, John Wiley & Sons, 2011

work page 2011

[36] [36]

Identiﬁcation an d control of dynamical systems using neural networks,

K. S. Narendra and K. Parthasarathy, “Identiﬁcation an d control of dynamical systems using neural networks,” IEEE Trans. Neural Netw. , vol. 1, no. 1, pp. 4-27, March 1990

work page 1990