Robust Methods for High-Dimensional Linear Learning

被引:0
作者
Merad, Ibrahim [1 ]
Gaiffas, Stephane [1 ,2 ]
机构
[1] Univ Paris Diderot, LPSM, UMR 8001, Paris, France
[2] Ecole Normale Super, DMA, Paris, France
关键词
robust methods; heavy-tailed data; outliers; sparse recovery; mirror descent; general; ization error; VARIABLE SELECTION; REGRESSION SHRINKAGE; ORACLE INEQUALITIES; LASSO; ESTIMATORS; SPARSITY; SLOPE; REGULARIZATION; RECOVERY; BOUNDS;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We propose statistically robust and computationally efficient linear learning methods in the highdimensional batch setting, where the number of features d may exceed the sample size n. We employ, in a generic learning setting, two algorithms depending on whether the considered loss function is gradient-Lipschitz or not. Then, we instantiate our framework on several applications including vanilla sparse, group-sparse and low-rank matrix recovery. This leads, for each application, to efficient and robust learning algorithms, that reach near-optimal estimation rates under heavy-tailed distributions and the presence of outliers. For vanilla s-sparsity, we are able to reach the s log(d)/n rate under heavy-tails and eta-corruption, at a computational cost comparable to that of non-robust analogs. We provide an efficient implementation of our algorithms in an open-source Python library called linlearn, by means of which we carry out numerical experiments which confirm our theoretical findings together with a comparison to other recent approaches proposed in the literature.
引用
收藏
页数:44
相关论文
共 91 条
  • [1] Tropp JA, 2015, Arxiv, DOI arXiv:1501.01571
  • [2] Agarwal A., 2012, Advances in Neural Information Processing Systems, V25
  • [3] The space complexity of approximating the frequency moments
    Alon, N
    Matias, Y
    Szegedy, M
    [J]. JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 1999, 58 (01) : 137 - 147
  • [4] [Anonymous], 2010, COLT 2010 23 C LEARN
  • [5] ROBUST LINEAR LEAST SQUARES REGRESSION
    Audibert, Jean-Yves
    Catoni, Olivier
    [J]. ANNALS OF STATISTICS, 2011, 39 (05) : 2766 - 2794
  • [6] Robust Linear Regression: Optimal Rates in Polynomial Time
    Bakshi, Ainesh
    Prasad, Adarsh
    [J]. STOC '21: PROCEEDINGS OF THE 53RD ANNUAL ACM SIGACT SYMPOSIUM ON THEORY OF COMPUTING, 2021, : 102 - 115
  • [7] Balakrishnan Sivaraman, 2017, C LEARN THEOR, P169
  • [8] SLOPE MEETS LASSO: IMPROVED ORACLE BOUNDS AND OPTIMALITY
    Bellec, Pierre C.
    Lecue, Guillaume
    Tsybakov, Alexandre B.
    [J]. ANNALS OF STATISTICS, 2018, 46 (6B) : 3603 - 3642
  • [9] SIMULTANEOUS ANALYSIS OF LASSO AND DANTZIG SELECTOR
    Bickel, Peter J.
    Ritov, Ya'acov
    Tsybakov, Alexandre B.
    [J]. ANNALS OF STATISTICS, 2009, 37 (04) : 1705 - 1732
  • [10] Normalized Iterative Hard Thresholding: Guaranteed Stability and Performance
    Blumensath, Thomas
    Davies, Mike E.
    [J]. IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING, 2010, 4 (02) : 298 - 309