Heavy Lasso: sparse penalized regression under heavy-tailed noise via data-augmented soft-thresholding

The Tien Mai

REVIEW 3 major objections 8 minor 32 references

Heavy Lasso: sparse penalized regression under heavy-tailed noise via data-augmented soft-thresholding

T0 review · 3 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Heavy Lasso swaps squared loss for Student-t loss to stay accurate under heavy-tailed noise.

desk verdict Plausible EM robust Lasso with solid numerics, but the main theorem relies on an LRSC lemma that is false under the stated assumptions. read the letter →

arxiv 2506.07790 v1 pith:TD6B2BXJ submitted 2025-06-09 stat.ME stat.ML

The Tien Mai This is my paper

classification stat.MEstat.ML MSC 62J0762F35

keywords robustregressionLassoheavy-tailednoiseStudent-tlosssoft-thresholdingdataaugmentationnon-asymptoticboundshigh-dimensional

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 proposes Heavy Lasso, a sparse linear regression method that replaces the classical squared loss with a Student-t-inspired loss, quadratic near zero and logarithmic for large residuals. The central claim is that this loss lets the Lasso estimator achieve the same non-asymptotic error rates as Huber-loss Lasso under only finite second-moment noise and standard high-dimensional design assumptions. A data-augmentation scheme rewrites the loss as a weighted squared loss, so the estimator can be computed by alternating weight updates with ordinary Lasso soft-thresholding. If the theory holds, Heavy Lasso offers a practical robust drop-in replacement for classical Lasso without extra tuning of a robustness threshold.

What carries the argument

The central object is the Student loss $\ell_i(\beta) = (\nu+1)\log(1 + (y_i - x_i^\top\beta)^2/\nu)$, which is quadratic for small residuals and logarithmic for large ones. The computational machinery is the scale-mixture representation of the Student-t distribution, which turns each iteration into a weighted Lasso problem with weights $\omega_i^{(t)} = (\nu+1)/(\nu + (y_i - x_i^\top\beta^{(t)})^2)$, solved by soft-thresholding. The theoretical machinery is the localized restricted strong convexity inequality (Lemma 2) that gives a quadratic lower bound on the loss increment in a restricted cone, together with the gradient concentration bound (Lemma 1) that controls the stochastic term.

What would settle it

Compute, for a design satisfying the restricted eigenvalue condition and heavy-tailed errors, the quantity $L(\beta^*+\Delta) - L(\beta^*) - \langle\nabla L(\beta^*), \Delta\rangle - \kappa\|\Delta\|_2^2$ over directions $\Delta$ with $\|\Delta_{S^c}\|_1 \le 3\|\Delta_S\|_1$; if this quantity is ever negative, or if a strictly positive tolerance must be subtracted to make it nonnegative, then the no-tolerance localized strong convexity needed for Theorem 1 fails.

Watch

Extended reading notes

Core claim

The central claim is that the Heavy Lasso estimator, defined by minimizing $$\frac{1}{2n}\sum_{i=1}^n (\nu+1)\log(1+(y_i-x_i^\top\$\beta$)^2/\nu) + \$\lambda$\|\$\beta$\|_1$$, satisfies, with high probability and for $\lambda \asymp \sqrt{\log p / n}$, the bounds $\|\hat{\beta}-\beta^*\|_2 \lesssim \sqrt{s\log p/n}$ and $\|\hat{\beta}-\beta^*\|_1 \lesssim s\sqrt{\log p/n}$, provided the design is bounded, the empirical Gram matrix satisfies the restricted eigenvalue condition, and the noise has finite second moment. The proof follows the localized convexity framework: a local restricted strong convexity inequality for the Student loss is combined with a Bernstein concentration bound on the gradient at the true parameter. The method is computationally tractable because the Student loss admits a scale-mixture representation, so each update reduces to solving a weighted Lasso via coordinate descent and soft-thresholding.

Load-bearing premise

The proof assumes the Student loss curves upward at least quadratically around the true coefficients, with no slack, on a restricted set of directions, and the paper cites this from earlier work without checking that this particular loss satisfies it; if the curvature inequality needs a tolerance or fails, the stated rates do not follow.

Editorial extensions

If this is right

Heavy Lasso matches the $\ell_2$ and $\ell_1$ error rates of Huber Lasso, so it is a theoretically grounded robust alternative to classical Lasso under heavy-tailed noise.
The rates hold for $p \gg n$ as long as $s\log p \ll n$, meaning the method is usable in genuinely high-dimensional regimes.
Because each iteration is a weighted Lasso, existing Lasso solvers can be reused with one extra data-augmentation step, lowering implementation cost.
The method requires only finite second moment of the noise, so it extends beyond sub-Gaussian assumptions.
Simulations indicate Heavy Lasso is competitive under Gaussian noise and sometimes clearly better under $t_3$ and Cauchy noise, especially in higher-dimensional settings.

Reading between the lines

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

The paper does not prove that the Student loss satisfies the localized restricted strong convexity condition with no tolerance term; verifying this condition, or showing it requires a tolerance, would settle whether the stated rates are real.
If the localized convexity inequality only holds with a tolerance term, the error bounds would likely degrade by a term depending on that tolerance, weakening the practical guarantee.
The scale-mixture formulation suggests an immediate extension to generalized linear models or to Bayesian treatments where the latent weights play the role of observation-specific noise scales.
The Cauchy-noise simulations, where Heavy Lasso is the only method with low estimation error in high dimensions, hint that the loss may remain useful even beyond its theoretical moment condition, but this would need separate justification.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Hi [Name],

Quick take on Mai's Heavy Lasso. The core idea is straightforward: take the Student-t robust loss, add an L1 penalty, fit via EM/soft-thresholding, and claim the same rates as Huber Lasso using Loh's localized convexity. That combination is mildly new (Cauchy loss is ν=1, and EM for t-regression is textbook), but it's a sensible thing to try, and the numerical section is solid: the method is competitive across Gaussian, t, and Cauchy noise, with code provided.

The problem is the theory. Theorem 1's proof leans entirely on Lemma 2, a borrowed LRSC claim. As stated, that lemma is false under Assumption 1. For L(β) = (ν+1)/(2n) Σ log(1+(y_i−x_i^T β)^2/ν), the Hessian at β* is (ν+1)/n Σ x_i x_i^T (ν−ε_i^2)/(ν+ε_i^2)^2. Since A3 only requires finite second moment, the coefficients can be negative. Concrete: ν=1, n=2, p=3, X rows (1,0.1,0) and (1,0,0.1), β*=(1,0,0), ε_i = ±√3. This satisfies A1–A3 and RE on s=1. Both residuals give factor (ν+1)(ν−ε^2)/(ν+ε^2)^2 = −1/4, so H_{11}(β*) = −1/4. Then L(β*+t e1)−L(β*)−⟨∇L(β*), t e1⟩ = −t^2/8 + O(t^3), which violates the κ‖Δ‖^2 lower bound for any κ>0. The proof's Step 2 has no foundation. The counterexample also shows β* need not be a local minimum, so this isn't a mere technicality.

There's a second, smaller gap: Lemma 1 uses E[ε/(ν+ε^2)] = 0, which follows from symmetry, not from the stated zero-mean assumption. And the tuning description switches from BIC to AIC between Section 3.3 and 4.1.

If the author can add a genuine curvature condition (e.g., bound the noise or prove a high-probability weighted-RE condition with the sign-changing weights), the result may be salvageable. As it stands, the theorem doesn't follow from the stated assumptions. Still, I'd send this to referees rather than desk-reject. The method is plausible, the experiments are honest, and the flaw is concrete enough to be fixable. It just needs real work on the RSC side.

Best, [Your name]

Referee Report

3 major / 8 minor

Summary. The paper introduces "Heavy Lasso," a penalized regression estimator that replaces the squared loss with the Student-t log loss ℓ_i(β) = (ν+1) log(1 + (y_i − x_i^Tβ)²/ν) inside an ℓ1-penalized objective. It claims non-asymptotic ℓ2 and ℓ1 error bounds of order √(s log p / n) and s√(log p / n) under bounded design, a restricted eigenvalue condition, and finite second moment of the noise. The estimator is computed by an EM-style data augmentation combined with coordinate-descent soft-thresholding updates, and the paper reports simulations across Gaussian, contaminated, t3, and Cauchy noise settings, a real-data application, and an R package.

Significance. If the theoretical claim were established, the result would be a useful addition to the robust high-dimensional regression literature: it would show that a bounded-influence, non-convex Student-type loss achieves the same rates as the Huber loss while remaining computationally simple via data augmentation and soft-thresholding. The paper has concrete strengths: extensive simulations across several noise settings, a real-data demonstration, and a public R implementation. However, the central theoretical guarantee is not currently established because the localized restricted strong convexity lemma used in the proof is false under the stated assumptions, and the stochastic gradient control requires an unstated symmetry condition. The paper would need substantive revision before the main claim is supported.

major comments (3)

[Appendix A, Lemma 2 and Step 2 of Theorem 1] The localized restricted strong convexity inequality (5), L(β*+Δ) − L(β*) − ⟨∇L(β*), Δ⟩ ≥ (κ/2)‖Δ‖₂², is asserted for the Student loss with no tolerance term, but it is false under Assumption 1. For L(β) = (ν+1)/(2n) Σ log(1 + (y_i − x_i^Tβ)²/ν), the Hessian at β* is H(β*) = (ν+1)/n Σ x_i x_i^T (ν − ε_i²)/(ν + ε_i²)². Assumption A3 only requires finite second moment, so the scalar coefficients can be negative. Concretely, take ν=1, n=2, p=3, x1=(1,0.1,0), x2=(1,0,0.1), β*=(1,0,0), s=1, and ε_i i.i.d. uniform on {±√3}; this satisfies A1–A3 and the RE condition on the s=1 cone, but H₁₁(β*) = −1/4. For Δ = t(1,0,0), the left side of (5) equals (1/2)H₁₁ t² + O(t³) = −t²/8 + O(t³), which contradicts (5) for any κ > 0 and arbitrarily small t. Thus Step 2 of the proof has no valid foundation, and Theorem 1 as stated does not follow. The authors need to verify the LRSC under an additional condition (e.g., |ε_i| bounded away from √ν) or introduce a tolerance term and track it through the entire argument.
[Appendix A, Lemma 1] Step 3 of the proof of Lemma 1 states that E[w_i] = 0 for w_i = ε_i/(ν + ε_i²), and the zero-mean assumption on ε_i is used to justify this. However, E[ε/(ν + ε²)] = 0 does not follow from E[ε] = 0; it requires symmetry of the noise distribution about zero. Without E[w_i] = 0, the Bernstein bound on ‖∇L(β*)‖∞ is not established, and the bound ‖∇L(β*)‖∞ ≤ λ/2 used in Step 3 of Theorem 1 has no basis. Either add a symmetry assumption to Assumption 1 or provide a different argument for controlling the gradient.
[Section 3.2, Algorithm 1] The soft-thresholding update in Algorithm 1 is missing the factor n. For the weighted Lasso objective (4), the coordinate-wise subproblem is (1/(2n)) Σ_i ω_i (r_ij − x_ij β_j)² + λ|β_j|, whose minimizer is β_j = (1/A_j) S(z_j, nλ), not β_j = (1/A_j) S(z_j, λ) as written. As stated, Algorithm 1 does not solve the optimization problem that defines the estimator. Please correct the update or state explicitly that the λ appearing in the algorithm is rescaled by n, and ensure consistency with the λ used in the theory and the tuning sections.

minor comments (8)

[Section 3.3 vs Section 4.1] Section 3.3 says the tuning parameter λ is selected by minimizing BIC and reports that AIC performs worse, while Section 4.1 states that the proposed Heavy Lasso method is tuned using AIC. Please reconcile this inconsistency and report which criterion was actually used in the simulations and real-data experiment.
[Section 4] The degrees of freedom ν of the Student loss is a free parameter, but the simulations do not report the value of ν used or any sensitivity analysis with respect to ν. Please specify the chosen ν and how it was selected.
[Theorem 1] The theorem states bounds "with high probability" without quantifying the failure probability or stating how the constants in the bounds depend on ν, K, κ, and the confidence level. Please state the result with explicit probability 1 − δ and explicit constants.
[Appendix A, Lemma 1] Lemma 1 says "there exist universal constants c1, c2 > 0" but only c1 appears in the bound, and the phrase "depending only on ν, , K" contains a typo. Please clean up the statement.
[Section 4.2.3, Tables 3 and 4] The Cauchy noise setting has infinite variance and therefore falls outside Assumption A3, so Theorem 1 does not cover the Cauchy simulations. Please state this limitation explicitly when discussing those results.
[Section 2.2] The claim that introducing a scaling constant c > 0 in the loss "does not alter the theoretical properties" is informal; either prove that the theorem remains valid for the scaled loss or qualify the claim.
[Section 5] The real-data experiment does not report how λ and ν were chosen for the proposed method. Please include these details for reproducibility.
[Title and text] There is a typo in the abstract/header "heavy-tail ed" and the GitHub URL in Section 3.3 is truncated as "heavyla"; both should be corrected.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the proof applies an externally cited localized-convexity framework, and no fitted constants or author-specific uniqueness results drive the conclusion.

full rationale

The Heavy Lasso derivation is an application of the Loh (2017, 2021) localized restricted strong convexity framework to a Student-t loss. Step 2 of the proof of Theorem 1 invokes Lemma 2, explicitly attributed to Appendix B.3 of Loh (2021), and the paper's own Lemma 1 gives the gradient bound using Bernstein's inequality. The regularization scale and rates are not fitted to data, and the claimed rates are standard consequences of the cited LRSC inequality. The author's self-citations (Mai 2024, 2025a, 2025b) appear only in the concluding discussion of future Bayesian extensions and are not load-bearing for the main theorem or algorithm. No ansatz is smuggled in via self-citation: the Student loss is defined in the paper, and its curvature property is imported from an external source, not from the author's prior work. The skeptic's concern that Lemma 2 may fail for the stated noise assumptions is a possible correctness or verification gap, not a circularity: a false external lemma would invalidate Theorem 1 rather than make it equivalent to its input. Since the central derivation reduces neither to a fitted quantity nor to a self-citation chain, the circularity score is 1.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard high-dimensional assumptions (RE condition, bounded design, finite second moment), an unstated symmetry assumption for the gradient bound, and a borrowed localized strong convexity lemma that is not re-verified for the t-loss.

free parameters (2)

degrees of freedom ν
User-specified tuning parameter controlling the robustness/curvature trade-off; the paper gives no data-driven rule for choosing it.
regularization parameter λ
Set theoretically to √(log p/n) for the rates, but in simulations selected by BIC or AIC (the paper is inconsistent about which).

assumptions (4)

domain assumption Restricted eigenvalue (RE) condition on the empirical Gram matrix (Assumption 1, A2)
Standard in high-dimensional Lasso theory; ensures identifiability of the sparse vector.
domain assumption Noise has finite second moment (Assumption 1, A3)
Needed for the concentration argument in Lemma 1; the method is still applied to Cauchy noise in simulations where this fails.
ad hoc to paper Noise distribution is symmetric so that E[ε/(ν+ε²)] = 0
Implicitly used in Lemma 1 to center the gradient; the paper only states zero mean, which is not sufficient for this expectation to vanish.
domain assumption The Student loss satisfies the localized restricted strong convexity inequality (Lemma 2, cited from Loh 2021)
The theorem's rates depend on this curvature condition; the paper does not verify that the t-loss meets the conditions of the cited lemma.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Heavy Lasso: sparse penalized regression under heavy-tailed noise via data-augmented soft-thresholding." pith.science (2026). https://pith.science/paper/TD6B2BXJ

@misc{pith2026250607790,
  author       = {Pith},
  title        = {Pith review of: Heavy Lasso: sparse penalized regression under heavy-tailed noise via data-augmented soft-thresholding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TD6B2BXJ}},
  note         = {Machine review of arXiv:2506.07790}
}

abstract

High-dimensional linear regression is a fundamental tool in modern statistics, particularly when the number of predictors exceeds the sample size. The classical Lasso, which relies on the squared loss, performs well under Gaussian noise assumptions but often deteriorates in the presence of heavy-tailed errors or outliers commonly encountered in real data applications such as genomics, finance, and signal processing. To address these challenges, we propose a novel robust regression method, termed Heavy Lasso, which incorporates a loss function inspired by the Student's t-distribution within a Lasso penalization framework. This loss retains the desirable quadratic behavior for small residuals while adaptively downweighting large deviations, thus enhancing robustness to heavy-tailed noise and outliers. Heavy Lasso enjoys computationally efficient by leveraging a data augmentation scheme and a soft-thresholding algorithm, which integrate seamlessly with classical Lasso solvers. Theoretically, we establish non-asymptotic bounds under both $\ell_1$ and $\ell_2 $ norms, by employing the framework of localized convexity, showing that the Heavy Lasso estimator achieves rates comparable to those of the Huber loss. Extensive numerical studies demonstrate Heavy Lasso's superior performance over classical Lasso and other robust variants, highlighting its effectiveness in challenging noisy settings. Our method is implemented in the R package heavylasso available on Github.

Figures

Figures reproduced from arXiv: 2506.07790 by the authors.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

[1]

Alfons, A. (2021). robustHD: An R package for robust regres sion with high-dimensional data. Journal of Open Source Software , 6(67):3786

work page 2021
[2]

C., Lecu´ e, G., and Tsybakov, A

Bellec, P. C., Lecu´ e, G., and Tsybakov, A. B. (2018). Slope meet s lasso: improved oracle bounds and optimality. The Annals of Statistics , 46(6B):3603–3642

work page 2018
[3]

Bhatia, K., Jain, P., and Kar, P. (2015). Robust regression via ha rd thresholding. Advances in neural information processing systems , 28

work page 2015
[4]

and Van De Geer, S

B¨ uhlmann, P. and Van De Geer, S. (2011). Statistics for high-dimensional data: methods, theory and applications. Springer Science & Business Media

work page 2011
[5]

Bunea, F., Tsybakov, A., and Wegkamp, M. (2007). Sparsity ora cle inequalities for the lasso. Electronic Journal of Statistics , 1:169–194

work page 2007
[6]

Chang, L., Roberts, S., and Welsh, A. (2018). Robust lasso regr ession using tukey’s biweight criterion. Technometrics, 60(1):36–47

work page 2018
[7]

and Thompson, P

Dalalyan, A. and Thompson, P. (2019). Outlier-robust estimatio n of a sparse linear model using ℓ1-penalized Huber’s M -estimator. Advances in neural information processing systems , 32

work page 2019
[8]

Friedman, J., Hastie, T., and Tibshirani, R. (2010). Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of Statistical Software , 33(1):1–22

work page 2010

Show all 32 references

[9]

and Majumdar, S

Ghosh, A. and Majumdar, S. (2020). Ultrahigh-dimensional rob ust and eﬃcient sparse regres- sion using non-concave penalized density power divergence. IEEE Transactions on Information Theory, 66(12):7812–7827

2020
[10]

Giraud, C. (2021). Introduction to high-dimensional statistics . Chapman and Hall/CRC

2021
[11]

H., and Friedman, J

Hastie, T., Tibshirani, R., Friedman, J. H., and Friedman, J. H. (20 09). The elements of statistical learning: data mining, inference, and predict ion, volume 2. Springer
[12]

and Lerasle, M

Lecu´ e, G. and Lerasle, M. (2020). Robust machine learning by median-of-means: theory and practice. The Annals of Statistics , 48(2):906–931

2020
[13]

Loh, P.-L. (2017). Statistical consistency and asymptotic no rmality for high-dimensional robust m-estimators. The Annals of Statistics , 45(2):866

2017
[14]

Loh, P.-L. (2021). Scale calibration for high-dimensional robus t regression. Electronic Journal of Statistics , 15(2):5933–5994

2021
[15]

Loh, P.-L. (2024). A theoretical review of modern robust sta tistics. Annual Review of Statistics and Its Application , 12

2024
[16]

Lounici, K. (2008). Sup-norm convergence rate and sign conc entration property of lasso and dantzig estimators. Electronic Journal of Statistics , 2:90–102

2008
[17]

Mai, T. T. (2024). Concentration of a Sparse Bayesian Model W ith Horseshoe Prior in Esti- mating High-Dimensional Precision Matrix. Stat, 13(4):e70008. 17

2024
[18]

Mai, T. T. (2025a). Concentration properties of fractional p osterior in 1-bit matrix completion. Machine Learning, 114(1):7

2025
[19]

Mai, T. T. (2025b). High-dimensional Bayesian Tobit regression for censored response with Horseshoe prior. arXiv preprint arXiv:2505.08288

2025 arXiv
[20]

C., Sunshine, M., Liu, H., Varma, S., Kohn, K

Reinhold, W. C., Sunshine, M., Liu, H., Varma, S., Kohn, K. W., Morris , J., Doroshow, J., and Pommier, Y. (2012). Cellminer: a web-based suite of genomic and pha rmacologic tools to explore transcript and drug patterns in the nci-60 cell line set. Cancer research, 72(14):3499–3511

2012
[21]

and Bogdan, M

Rejchel, W. and Bogdan, M. (2020). Rank-based lasso-eﬃcien t methods for high-dimensional robust model selection. Journal of Machine Learning Research , 21(244):1–47

2020
[22]

and Zhu, J

Rosset, S. and Zhu, J. (2007). Piecewise linear regularized solu tion paths. Annals of statistics , 35(3):1012–1030

2007
[23]

and Yohai, V

Smucler, E. and Yohai, V. J. (2017). Robust and sparse estima tors for linear regression models. Computational Statistics & Data Analysis , 111:116–130

2017
[24]

Tibshirani, R. (1996). Regression shrinkage and selection via th e lasso. Journal of the Royal Statistical Society Series B: Statistical Methodology , 58(1):267–288

1996
[25]

Wang, H., Li, G., and Jiang, G. (2007). Robust regression shrink age and consistent variable selection through the lad-lasso. Journal of Business & Economic Statistics , 25(3):347–355

2007
[26]

Wang, L., Peng, B., Bradic, J., Li, R., and Wu, Y. (2020). A tuning- free robust and eﬃ- cient approach to high-dimensional regression. Journal of the American Statistical Association , 115(532):1700–1714

2020
[27]

and Karunamuni, R

Wang, Y. and Karunamuni, R. J. (2022). High-dimensional robu st regression with lq-loss functions. Computational Statistics & Data Analysis , 176:107567

2022
[28]

and Huang, J

Yi, C. and Huang, J. (2017). Semismooth newton coordinate de scent algorithm for elastic-net penalized huber loss regression and quantile regression. Journal of Computational and Graphical Statistics, 26(3):547–557

2017
[29]

Zhang, C.-H. (2010). Nearly unbiased variable selection under m inimax concave penalty. The Annals of Statistics , 38(2):894

2010
[30]

and Yu, B

Zhao, P. and Yu, B. (2006). On model selection consistency of lasso. Journal of Machine learning research, 7(Nov):2541–2563

2006
[31]

Zhou, L., Wang, B., and Zou, H. (2024). Sparse convoluted ran k regression in high dimensions. Journal of the American Statistical Association , 119(546):1500–1512

2024
[32]

Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American statistical association, 101(476):1418–1429. 18

2006

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Alfons, A. (2021). robustHD: An R package for robust regres sion with high-dimensional data. Journal of Open Source Software , 6(67):3786

work page 2021

[2] [2]

C., Lecu´ e, G., and Tsybakov, A

Bellec, P. C., Lecu´ e, G., and Tsybakov, A. B. (2018). Slope meet s lasso: improved oracle bounds and optimality. The Annals of Statistics , 46(6B):3603–3642

work page 2018

[3] [3]

Bhatia, K., Jain, P., and Kar, P. (2015). Robust regression via ha rd thresholding. Advances in neural information processing systems , 28

work page 2015

[4] [4]

and Van De Geer, S

B¨ uhlmann, P. and Van De Geer, S. (2011). Statistics for high-dimensional data: methods, theory and applications. Springer Science & Business Media

work page 2011

[5] [5]

Bunea, F., Tsybakov, A., and Wegkamp, M. (2007). Sparsity ora cle inequalities for the lasso. Electronic Journal of Statistics , 1:169–194

work page 2007

[6] [6]

Chang, L., Roberts, S., and Welsh, A. (2018). Robust lasso regr ession using tukey’s biweight criterion. Technometrics, 60(1):36–47

work page 2018

[7] [7]

and Thompson, P

Dalalyan, A. and Thompson, P. (2019). Outlier-robust estimatio n of a sparse linear model using ℓ1-penalized Huber’s M -estimator. Advances in neural information processing systems , 32

work page 2019

[8] [8]

Friedman, J., Hastie, T., and Tibshirani, R. (2010). Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of Statistical Software , 33(1):1–22

work page 2010