Generalization bounds for sparse random feature expansions

Random feature methods have been successful in various machine learning tasks, are easy to compute, and come with theoretical accuracy bounds. They serve as an alternative approach to standard neural networks since they can represent similar function spaces without a costly training phase. However,...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Applied and computational harmonic analysis Ročník 62; s. 310 - 330
Hlavní autori: Hashemi, Abolfazl, Schaeffer, Hayden, Shi, Robert, Topcu, Ufuk, Tran, Giang, Ward, Rachel
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier Inc 01.01.2023
Predmet:
ISSN:1063-5203, 1096-603X
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract Random feature methods have been successful in various machine learning tasks, are easy to compute, and come with theoretical accuracy bounds. They serve as an alternative approach to standard neural networks since they can represent similar function spaces without a costly training phase. However, for accuracy, random feature methods require more measurements than trainable parameters, limiting their use for data-scarce applications. We introduce the sparse random feature expansion to obtain parsimonious random feature models. We leverage ideas from compressive sensing to generate random feature expansions with theoretical guarantees even in the data-scarce setting. We provide generalization bounds for functions in a certain class depending on the number of samples and the distribution of features. By introducing sparse features, i.e. features with random sparse weights, we provide improved bounds for low order functions. We show that our method outperforms shallow networks in several scientific machine learning tasks.
AbstractList Random feature methods have been successful in various machine learning tasks, are easy to compute, and come with theoretical accuracy bounds. They serve as an alternative approach to standard neural networks since they can represent similar function spaces without a costly training phase. However, for accuracy, random feature methods require more measurements than trainable parameters, limiting their use for data-scarce applications. We introduce the sparse random feature expansion to obtain parsimonious random feature models. We leverage ideas from compressive sensing to generate random feature expansions with theoretical guarantees even in the data-scarce setting. We provide generalization bounds for functions in a certain class depending on the number of samples and the distribution of features. By introducing sparse features, i.e. features with random sparse weights, we provide improved bounds for low order functions. We show that our method outperforms shallow networks in several scientific machine learning tasks.
Author Schaeffer, Hayden
Hashemi, Abolfazl
Shi, Robert
Topcu, Ufuk
Tran, Giang
Ward, Rachel
Author_xml – sequence: 1
  givenname: Abolfazl
  surname: Hashemi
  fullname: Hashemi, Abolfazl
  organization: Purdue University, United States of America
– sequence: 2
  givenname: Hayden
  surname: Schaeffer
  fullname: Schaeffer, Hayden
  email: hschaeff@andrew.cmu.edu
  organization: Carnegie Mellon University, United States of America
– sequence: 3
  givenname: Robert
  orcidid: 0000-0001-7850-8291
  surname: Shi
  fullname: Shi, Robert
  organization: The University of Texas at Austin, United States of America
– sequence: 4
  givenname: Ufuk
  surname: Topcu
  fullname: Topcu, Ufuk
  organization: The University of Texas at Austin, United States of America
– sequence: 5
  givenname: Giang
  surname: Tran
  fullname: Tran, Giang
  organization: University of Waterloo, Canada
– sequence: 6
  givenname: Rachel
  surname: Ward
  fullname: Ward, Rachel
  organization: The University of Texas at Austin, United States of America
BookMark eNp9kE1LAzEQhoNUsK3-AU_7B3adJPvRiBcpWoWCFwVvIR8TTGmTkmxF_fXuqicPPc0c5nl555mRSYgBCbmkUFGg7dWmUuZNVQwYq2BRAfATMqUg2rIF_joZ95aXDQN-RmY5bwAorRsxJdcrDJjU1n-p3sdQ6HgINhcupiLvVcpYJBVs3BUOVX9IWODHXoU8nOZzcurUNuPF35yTl_u75-VDuX5aPS5v16Xhdd2XtquRaaGFclwY7TiIRjPbsHpoiaztWmucFQBN54BrY00HTW07o7miqm34nCx-c02KOSd00vj-p22flN9KCnJ0IDdydCBHBxIWckgfUPYP3Se_U-nzOHTzC-Hw1LvHJLPxGAxan9D00kZ_DP8GpgN4eQ
CitedBy_id crossref_primary_10_1111_mice_13372
crossref_primary_10_1088_1361_6501_acea9e
crossref_primary_10_1007_s43670_023_00063_9
crossref_primary_10_1137_23M1624749
crossref_primary_10_1177_14759217251360635
crossref_primary_10_1615_Int_J_UncertaintyQuantification_2025051106
crossref_primary_10_1016_j_apacoust_2025_110562
crossref_primary_10_1214_25_AOS2492
crossref_primary_10_3390_atmos14091451
crossref_primary_10_1137_24M1648703
crossref_primary_10_1016_j_acha_2025_101789
crossref_primary_10_1093_sysbio_syae026
crossref_primary_10_1007_s10444_024_10147_2
crossref_primary_10_1109_TAP_2024_3501410
crossref_primary_10_1016_j_aei_2024_102394
crossref_primary_10_1016_j_ymssp_2023_111032
crossref_primary_10_1007_s43670_025_00100_9
crossref_primary_10_1109_TAC_2024_3486655
crossref_primary_10_1016_j_sigpro_2025_110239
crossref_primary_10_1016_j_ecolind_2024_112835
crossref_primary_10_1016_j_laa_2024_04_023
crossref_primary_10_1007_s11538_023_01174_z
Cites_doi 10.1016/j.acha.2015.02.003
10.1109/JSTSP.2007.910281
10.1137/130916138
10.1109/TIT.2006.871582
10.1016/j.cma.2017.07.038
10.1016/j.jcp.2015.09.001
10.1111/j.1467-9868.2009.00718.x
10.1109/18.256500
10.1111/j.1541-0420.2012.01751.x
10.1007/s10208-017-9350-3
10.1109/TIT.2006.885507
10.1016/j.jat.2012.01.008
10.1007/BF01397280
10.1007/s00365-010-9105-8
10.1137/080714488
10.1109/TIT.2009.2021377
10.1016/j.jcp.2011.01.002
10.1103/RevModPhys.34.123
10.1002/cpa.20124
10.1137/18M1194730
10.1090/S0025-5718-09-02319-9
10.1137/18M116798X
10.1007/s10208-012-9115-y
10.1016/j.jcss.2004.04.001
10.1080/10618600.2014.938812
10.1214/aos/1176348546
ContentType Journal Article
Copyright 2022 The Author(s)
Copyright_xml – notice: 2022 The Author(s)
DBID 6I.
AAFTH
AAYXX
CITATION
DOI 10.1016/j.acha.2022.08.003
DatabaseName ScienceDirect Open Access Titles
Elsevier:ScienceDirect:Open Access
CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
Mathematics
EISSN 1096-603X
EndPage 330
ExternalDocumentID 10_1016_j_acha_2022_08_003
S1063520322000653
GrantInformation_xml – fundername: MURI
  grantid: FA9550-19-1-0005
  funderid: https://doi.org/10.13039/100014036
– fundername: NSERC
  grantid: 2018-06135
  funderid: https://doi.org/10.13039/501100000038
– fundername: NSF
  grantid: DMS-1752116; DMS-1952735
  funderid: https://doi.org/10.13039/100000001
– fundername: MURI
  grantid: FA9550-21-1-0084
  funderid: https://doi.org/10.13039/100014036
GroupedDBID --K
--M
.~1
0R~
1B1
1RT
1~.
1~5
23M
4.4
457
4G.
5GY
5VS
6I.
7-5
71M
8P~
9JN
AACTN
AAEDT
AAEDW
AAFTH
AAIAV
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AASFE
AAXUO
ABAOU
ABFNM
ABJNI
ABMAC
ABVKL
ABXDB
ABYKQ
ACAZW
ACDAQ
ACGFS
ACRLP
ADBBV
ADEZE
ADFGL
ADMUD
AEBSH
AEKER
AENEX
AEXQZ
AFKWA
AFTJW
AGHFR
AGUBO
AGYEJ
AHHHB
AIEXJ
AIGVJ
AIKHN
AITUG
AJBFU
AJOXV
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
ARUGR
ASPBG
AVWKF
AXJTR
AZFZN
BKOJK
BLXMC
CAG
COF
CS3
DM4
EBS
EFBJH
EFLBG
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-2
G-Q
GBLVA
HVGLF
HZ~
IHE
IXB
J1W
KOM
LG5
M26
M41
MCRUF
MHUIS
MO0
N9A
NCXOZ
O-L
O9-
OAUVE
OK1
OZT
P-8
P-9
P2P
PC.
Q38
R2-
RIG
ROL
RPZ
SDF
SDG
SDP
SES
SEW
SPC
SPCBC
SSW
SSZ
T5K
WUQ
XPP
ZMT
~G-
9DU
AATTM
AAXKI
AAYWO
AAYXX
ABWVN
ACLOT
ACRPL
ACVFH
ADCNI
ADNMO
ADVLN
AEIPS
AEUPX
AFJKZ
AFPUW
AGQPQ
AIGII
AIIUN
AKBMS
AKRWK
AKYEP
ANKPU
APXCP
CITATION
EFKBS
~HD
ID FETCH-LOGICAL-c344t-d74e2b9b9af39cbf3095b2d524003e2676dcfd90057f03bcdc7054d7cb3a1a653
ISICitedReferencesCount 31
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000870987200003&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1063-5203
IngestDate Sat Nov 29 07:10:32 EST 2025
Tue Nov 18 21:55:19 EST 2025
Fri Feb 23 02:39:36 EST 2024
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Keywords Sparse optimization
Compressive sensing
Random features
60B20
Generalization error
65D15
68Q32
46N10
Language English
License This is an open access article under the CC BY license.
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c344t-d74e2b9b9af39cbf3095b2d524003e2676dcfd90057f03bcdc7054d7cb3a1a653
ORCID 0000-0001-7850-8291
OpenAccessLink https://dx.doi.org/10.1016/j.acha.2022.08.003
PageCount 21
ParticipantIDs crossref_citationtrail_10_1016_j_acha_2022_08_003
crossref_primary_10_1016_j_acha_2022_08_003
elsevier_sciencedirect_doi_10_1016_j_acha_2022_08_003
PublicationCentury 2000
PublicationDate January 2023
2023-01-00
PublicationDateYYYYMMDD 2023-01-01
PublicationDate_xml – month: 01
  year: 2023
  text: January 2023
PublicationDecade 2020
PublicationTitle Applied and computational harmonic analysis
PublicationYear 2023
Publisher Elsevier Inc
Publisher_xml – name: Elsevier Inc
References Schaeffer, Tran, Ward, Zhang (br0470) 2020; 18
Adcock, Brugiapaglia, Webster (br0010) 2017
Rosset, Swirszcz, Srebro, Zhu (br0570) 2007
Mei, Montanari (br0350) 2020
Guo, Zhu, Chow, Ibrahim (br0550) 2012
Kandasamy, Yu (br0610) 2016
Candes, Tao (br0090) 2006; 52
van Den Berg, Friedlander (br0680) 2009
Potts, Schmischke (br0390) 2019
Constantine, Emory, Larsson, Iaccarino (br0160) 2015; 302
Rauhut, Ward (br0430) 2012; 164
Arora, Du, Hu, Li, Wang (br0030) 2019
Potts, Schmischke (br0380) 2019
Barron (br0040) 1993; 39
Szabó, Sriperumbudur (br0670) 2019
D. Dua, C. Graff, UCI machine learning repository, 2017.
Rahimi, Recht (br0410) 2008
Ullah, Mianjy, Marinov, Arora (br0660) 2018
Du, Zhai, Poczos, Singh (br0200) 2018
Figueiredo, Nowak, Wright (br0600) 2007
Chkifa, Dexter, Tran, Webster (br0130) 2018; 311
Moosmann, Triggs, Jurie (br0360) 2006
Maass, Markram (br0330) 2004; 69
Cai, Xu, Zhang (br0070) 2009; 55
DeVore, Petrova, Wojtaszczyk (br0170) 2011; 33
Shi, Lee, Wahba (br0540) 2007
Jones (br0280) 1992; 20
Du, Hu, Kakade, Lee, Lei (br0590) 2020
Rauhut, Ward (br0440) 2016; 40
Donoho (br0180) 2006; 52
Rakotomamonjy, Flamary, Yger (br0580) 2013
Marelli, Sudret (br0340) 2014
Jacot, Gabriel, Hongler (br0270) 2018
Sriperumbudur, Szabó (br0490) 2015
Chen, Wang, Deng, Huang (br0630) 2017
Rudi, Rosasco (br0650) 2017
Ravikumar, Lafferty, Liu, Wasserman (br0640) 2009
Chitta, Jin, Jain (br0120) 2012
Candes, Romberg, Tao (br0080) 2006; 59
Yang, Li, Mahdavi, Jin, Zhou (br0520) 2012; 25
Sutherland, Schneider (br0500) 2015
Schaeffer, Tran, Ward (br0460) 2018; 78
Constantine, Eftekhari, Hokanson, Ward (br0150) 2017; 326
Heisenberg (br0250) 1927
Constantine, Dow, Wang (br0140) 2014; 36
Saltelli, Ratto, Andres, Campolongo, Cariboni, Gatelli, Saisana, Tarantola (br0450) 2008
Li, Ton, Oglic, Sejdinovic (br0320) 2019
Lim, Hastie (br0560) 2015
Harris (br0240) 2019
Rahimi, Recht (br0420) 2007
Fornasier, Schnass, Vybiral (br0220) 2012; 12
Rahimi, Recht (br0400) 2008
Yen, Lin, Lin, Ravikumar, Dhillon (br0530) 2014
Liu, Chen, Huang (br0620) 2020
Block (br0060) 1962; 34
Foucart, Rauhut (br0230) 2013
Kuo, Sloan, Wasilkowski, Woźniakowski (br0290) 2010; 79
Adcock (br0020) 2018; 18
Doostan, Owhadi (br0190) 2011; 230
Li, Liang (br0310) 2018
Foucart (10.1016/j.acha.2022.08.003_br0230) 2013
Chitta (10.1016/j.acha.2022.08.003_br0120) 2012
Fornasier (10.1016/j.acha.2022.08.003_br0220) 2012; 12
Ravikumar (10.1016/j.acha.2022.08.003_br0640) 2009
DeVore (10.1016/j.acha.2022.08.003_br0170) 2011; 33
Chen (10.1016/j.acha.2022.08.003_br0630) 2017
Adcock (10.1016/j.acha.2022.08.003_br0010) 2017
Candes (10.1016/j.acha.2022.08.003_br0080) 2006; 59
Schaeffer (10.1016/j.acha.2022.08.003_br0470) 2020; 18
Ullah (10.1016/j.acha.2022.08.003_br0660) 2018
Rakotomamonjy (10.1016/j.acha.2022.08.003_br0580) 2013
Rosset (10.1016/j.acha.2022.08.003_br0570) 2007
Li (10.1016/j.acha.2022.08.003_br0320) 2019
Potts (10.1016/j.acha.2022.08.003_br0380)
Constantine (10.1016/j.acha.2022.08.003_br0160) 2015; 302
Doostan (10.1016/j.acha.2022.08.003_br0190) 2011; 230
Marelli (10.1016/j.acha.2022.08.003_br0340) 2014
Cai (10.1016/j.acha.2022.08.003_br0070) 2009; 55
Donoho (10.1016/j.acha.2022.08.003_br0180) 2006; 52
Barron (10.1016/j.acha.2022.08.003_br0040) 1993; 39
Harris (10.1016/j.acha.2022.08.003_br0240)
Rudi (10.1016/j.acha.2022.08.003_br0650) 2017
Du (10.1016/j.acha.2022.08.003_br0590)
Figueiredo (10.1016/j.acha.2022.08.003_br0600) 2007
Rauhut (10.1016/j.acha.2022.08.003_br0430) 2012; 164
Adcock (10.1016/j.acha.2022.08.003_br0020) 2018; 18
Yang (10.1016/j.acha.2022.08.003_br0520) 2012; 25
van Den Berg (10.1016/j.acha.2022.08.003_br0680) 2009
Szabó (10.1016/j.acha.2022.08.003_br0670) 2019
Constantine (10.1016/j.acha.2022.08.003_br0150) 2017; 326
Rahimi (10.1016/j.acha.2022.08.003_br0410) 2008
Sutherland (10.1016/j.acha.2022.08.003_br0500)
Rahimi (10.1016/j.acha.2022.08.003_br0400) 2008
Lim (10.1016/j.acha.2022.08.003_br0560) 2015
Maass (10.1016/j.acha.2022.08.003_br0330) 2004; 69
Rahimi (10.1016/j.acha.2022.08.003_br0420) 2007
10.1016/j.acha.2022.08.003_br0210
Yen (10.1016/j.acha.2022.08.003_br0530) 2014
Du (10.1016/j.acha.2022.08.003_br0200)
Sriperumbudur (10.1016/j.acha.2022.08.003_br0490)
Block (10.1016/j.acha.2022.08.003_br0060) 1962; 34
Candes (10.1016/j.acha.2022.08.003_br0090) 2006; 52
Li (10.1016/j.acha.2022.08.003_br0310)
Liu (10.1016/j.acha.2022.08.003_br0620) 2020
Mei (10.1016/j.acha.2022.08.003_br0350)
Shi (10.1016/j.acha.2022.08.003_br0540) 2007
Arora (10.1016/j.acha.2022.08.003_br0030) 2019
Kuo (10.1016/j.acha.2022.08.003_br0290) 2010; 79
Potts (10.1016/j.acha.2022.08.003_br0390)
Moosmann (10.1016/j.acha.2022.08.003_br0360) 2006
Schaeffer (10.1016/j.acha.2022.08.003_br0460) 2018; 78
Heisenberg (10.1016/j.acha.2022.08.003_br0250) 1927
Constantine (10.1016/j.acha.2022.08.003_br0140) 2014; 36
Saltelli (10.1016/j.acha.2022.08.003_br0450) 2008
Jones (10.1016/j.acha.2022.08.003_br0280) 1992; 20
Kandasamy (10.1016/j.acha.2022.08.003_br0610) 2016
Chkifa (10.1016/j.acha.2022.08.003_br0130) 2018; 311
Guo (10.1016/j.acha.2022.08.003_br0550) 2012
Jacot (10.1016/j.acha.2022.08.003_br0270)
Rauhut (10.1016/j.acha.2022.08.003_br0440) 2016; 40
References_xml – year: 2015
  ident: br0500
  article-title: On the error of random Fourier features
– volume: 34
  start-page: 123
  year: 1962
  ident: br0060
  article-title: The perceptron: a model for brain functioning. I
  publication-title: Rev. Mod. Phys.
– start-page: 197
  year: 2017
  end-page: 207
  ident: br0630
  article-title: Group sparse additive machine
  publication-title: Proceedings of the 31st International Conference on Neural Information Processing Systems
– volume: 36
  start-page: A1500
  year: 2014
  end-page: A1524
  ident: br0140
  article-title: Active subspace methods in theory and practice: applications to Kriging surfaces
  publication-title: SIAM J. Sci. Comput.
– start-page: 1009
  year: 2009
  end-page: 1030
  ident: br0640
  article-title: Sparse additive models
  publication-title: J. R. Stat. Soc., Ser. B, Stat. Methodol.
– start-page: 890
  year: 2009
  end-page: 912
  ident: br0680
  article-title: Probing the Pareto frontier for basis pursuit solutions
  publication-title: SIAM J. Sci. Comput.
– start-page: 1
  year: 2007
  end-page: 5
  ident: br0540
  article-title: Detecting disease-causing genes by LASSO-Patternsearch algorithm
  publication-title: BMC Proc.
– start-page: 35
  year: 2017
  end-page: 3225
  ident: br0650
  article-title: Generalization properties of learning with random features
  publication-title: Proceedings of the 31st International Conference on Neural Information Processing Systems
– volume: 12
  start-page: 229
  year: 2012
  end-page: 262
  ident: br0220
  article-title: Learning functions of few arbitrary linear parameters in high dimensions
  publication-title: Found. Comput. Math.
– volume: 52
  start-page: 5406
  year: 2006
  end-page: 5425
  ident: br0090
  article-title: Near-optimal signal recovery from random projections: universal encoding strategies?
  publication-title: IEEE Trans. Inf. Theory
– start-page: 627
  year: 2015
  end-page: 654
  ident: br0560
  article-title: Learning interactions via hierarchical group-lasso regularization
  publication-title: J. Comput. Graph. Stat.
– volume: 302
  start-page: 1
  year: 2015
  end-page: 20
  ident: br0160
  article-title: Exploiting active subspaces to quantify uncertainty in the numerical simulation of the HyShot II scramjet
  publication-title: J. Comput. Phys.
– start-page: 2554
  year: 2014
  end-page: 2563
  ident: br0340
  article-title: UQlab: a framework for uncertainty quantification in Matlab
  publication-title: Vulnerability, Uncertainty, and Risk: Quantification, Mitigation, and Management
– year: 2008
  ident: br0450
  article-title: Global Sensitivity Analysis: the Primer
– volume: 25
  start-page: 476
  year: 2012
  end-page: 484
  ident: br0520
  article-title: Nyström method vs random Fourier features: a theoretical and empirical comparison
  publication-title: Adv. Neural Inf. Process. Syst.
– year: 2019
  ident: br0380
  article-title: Approximation of high-dimensional periodic functions with Fourier-based methods
– start-page: 1313
  year: 2008
  end-page: 1320
  ident: br0410
  article-title: Weighted sums of random kitchen sinks: replacing minimization with randomization in learning
  publication-title: Adv. Neural Inf. Process. Syst.
– year: 2019
  ident: br0390
  article-title: Learning multivariate functions with low-dimensional structures using polynomial bases
– start-page: 1
  year: 2007
  end-page: 10
  ident: br0420
  article-title: Random features for large-scale kernel machines
  publication-title: NIPS
– volume: 20
  start-page: 608
  year: 1992
  end-page: 613
  ident: br0280
  article-title: A simple lemma on greedy approximation in Hilbert space and convergence rates for projection pursuit regression and neural network training
  publication-title: Ann. Stat.
– volume: 69
  start-page: 593
  year: 2004
  end-page: 616
  ident: br0330
  article-title: On the computational power of circuits of spiking neurons
  publication-title: J. Comput. Syst. Sci.
– start-page: 2456
  year: 2014
  end-page: 2464
  ident: br0530
  article-title: Sparse random feature algorithm as coordinate descent in Hilbert space
  publication-title: Adv. Neural Inf. Process. Syst.
– volume: 59
  start-page: 1207
  year: 2006
  end-page: 1223
  ident: br0080
  article-title: Stable signal recovery from incomplete and inaccurate measurements
  publication-title: Commun. Pure Appl. Math.
– year: 2019
  ident: br0240
  article-title: Additive function approximation in the brain
– year: 2018
  ident: br0310
  article-title: Learning overparameterized neural networks via stochastic gradient descent on structured data
– reference: D. Dua, C. Graff, UCI machine learning repository, 2017.
– volume: 33
  start-page: 125
  year: 2011
  end-page: 143
  ident: br0170
  article-title: Approximation of functions of few variables in high dimensions
  publication-title: Constr. Approx.
– volume: 230
  start-page: 3015
  year: 2011
  end-page: 3034
  ident: br0190
  article-title: A non-adapted sparse approximation of PDEs with stochastic inputs
  publication-title: J. Comput. Phys.
– year: 2020
  ident: br0590
  article-title: Few-shot learning via learning the representation, provably
– volume: 164
  start-page: 517
  year: 2012
  end-page: 533
  ident: br0430
  article-title: Sparse Legendre expansions via l1-minimization
  publication-title: J. Approx. Theory
– year: 2020
  ident: br0350
  article-title: The generalization error of random features regression: precise asymptotics and double descent curve
– start-page: 827
  year: 2019
  end-page: 836
  ident: br0670
  article-title: On kernel derivative approximation with random Fourier features
  publication-title: Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics
– year: 2006
  ident: br0360
  article-title: Randomized clustering forests for building fast and discriminative visual vocabularies
  publication-title: NIPS
– volume: 18
  start-page: 1435
  year: 2020
  end-page: 1461
  ident: br0470
  article-title: Extracting structured dynamical systems using sparse optimization with very few samples
  publication-title: Multiscale Model. Simul.
– volume: 40
  start-page: 3
  year: 2016
  end-page: 351
  ident: br0440
  article-title: Interpolation via weighted l1 minimization
  publication-title: Appl. Comput. Harmon. Anal.
– start-page: 93
  year: 2017
  end-page: 124
  ident: br0010
  article-title: Compressed sensing approaches for polynomial approximation of high-dimensional functions
  publication-title: Compressed Sensing and its Applications
– year: 2018
  ident: br0200
  article-title: Gradient descent provably optimizes over-parameterized neural networks
– start-page: 69
  year: 2016
  end-page: 78
  ident: br0610
  article-title: Additive approximations in high dimensional nonparametric regression via the SALSA
  publication-title: International Conference on Machine Learning
– volume: 55
  start-page: 3388
  year: 2009
  end-page: 3397
  ident: br0070
  article-title: On recovery of sparse signals via
  publication-title: IEEE Trans. Inf. Theory
– year: 2013
  ident: br0230
  article-title: A Mathematical Introduction to Compressive Sensing
– start-page: 7322
  year: 2018
  end-page: 7332
  ident: br0660
  article-title: Streaming kernel PCA with
  publication-title: Features Proceedings of the 32nd International Conference on Neural Information Processing System
– start-page: 3905
  year: 2019
  end-page: 3914
  ident: br0320
  article-title: Towards a unified analysis of random Fourier features
  publication-title: International Conference on Machine Learning
– volume: 78
  start-page: 3279
  year: 2018
  end-page: 3295
  ident: br0460
  article-title: Extracting sparse high-dimensional dynamics from limited data
  publication-title: SIAM J. Appl. Math.
– start-page: 555
  year: 2008
  end-page: 561
  ident: br0400
  article-title: Uniform approximation of functions with random bases
  publication-title: 2008 46th Annual Allerton Conference on Communication, Control, and Computing
– volume: 326
  start-page: 402
  year: 2017
  end-page: 404
  ident: br0150
  article-title: A near-stationary subspace for ridge approximation
  publication-title: Comput. Methods Appl. Mech. Eng.
– volume: 79
  start-page: 953
  year: 2010
  end-page: 966
  ident: br0290
  article-title: On decompositions of multivariate functions
  publication-title: Math. Comput.
– start-page: 586
  year: 2007
  end-page: 597
  ident: br0600
  article-title: Gradient projection for sparse reconstruction: application to compressed sensing and other inverse problems
  publication-title: IEEE J. Sel. Top. Signal Process.
– volume: 18
  start-page: 661
  year: 2018
  end-page: 701
  ident: br0020
  article-title: Infinite-dimensional compressed sensing and function interpolation
  publication-title: Found. Comput. Math.
– start-page: 544
  year: 2007
  end-page: 558
  ident: br0570
  article-title: l-1 regularization in infinite dimensional feature spaces
  publication-title: International Conference on Computational Learning Theory
– start-page: 322
  year: 2019
  end-page: 332
  ident: br0030
  article-title: Fine-grained analysis of optimization and generalization for overparameterized two-layer neural networks
  publication-title: International Conference on Machine Learning
– start-page: 43
  year: 2013
  end-page: 66
  ident: br0580
  article-title: Learning with infinitely many features
  publication-title: Machine Learning
– volume: 311
  start-page: 1415
  year: 2018
  end-page: 1450
  ident: br0130
  article-title: Polynomial approximation via compressed sensing of high-dimensional functions on lower sets
  publication-title: Math. Comput.
– volume: 39
  start-page: 930
  year: 1993
  end-page: 945
  ident: br0040
  article-title: Universal approximation bounds for superpositions of a sigmoidal function
  publication-title: IEEE Trans. Inf. Theory
– start-page: 6194
  year: 2020
  end-page: 6204
  ident: br0620
  article-title: Sparse shrunk additive models
  publication-title: International Conference on Machine Learning
– volume: 52
  start-page: 1289
  year: 2006
  end-page: 1306
  ident: br0180
  article-title: Compressed sensing
  publication-title: IEEE Trans. Inf. Theory
– start-page: 161
  year: 2012
  end-page: 170
  ident: br0120
  article-title: Efficient kernel clustering using random Fourier features
  publication-title: 2012 IEEE 12th International Conference on Data Mining
– year: 2015
  ident: br0490
  article-title: Optimal rates for random Fourier features
– start-page: 567
  year: 2012
  end-page: 577
  ident: br0550
  article-title: Bayesian lasso for semiparametric structural equation models
  publication-title: Biometrics
– start-page: 172
  year: 1927
  end-page: 198
  ident: br0250
  article-title: About the descriptive content of quantum theoretical kinematics and mechanics
  publication-title: Z. Phys.
– year: 2018
  ident: br0270
  article-title: Neural tangent kernel: convergence and generalization in neural networks
– ident: 10.1016/j.acha.2022.08.003_br0350
– start-page: 69
  year: 2016
  ident: 10.1016/j.acha.2022.08.003_br0610
  article-title: Additive approximations in high dimensional nonparametric regression via the SALSA
– start-page: 1
  year: 2007
  ident: 10.1016/j.acha.2022.08.003_br0420
  article-title: Random features for large-scale kernel machines
– volume: 40
  start-page: 3
  issue: 2
  year: 2016
  ident: 10.1016/j.acha.2022.08.003_br0440
  article-title: Interpolation via weighted l1 minimization
  publication-title: Appl. Comput. Harmon. Anal.
  doi: 10.1016/j.acha.2015.02.003
– start-page: 586
  year: 2007
  ident: 10.1016/j.acha.2022.08.003_br0600
  article-title: Gradient projection for sparse reconstruction: application to compressed sensing and other inverse problems
  publication-title: IEEE J. Sel. Top. Signal Process.
  doi: 10.1109/JSTSP.2007.910281
– volume: 25
  start-page: 476
  year: 2012
  ident: 10.1016/j.acha.2022.08.003_br0520
  article-title: Nyström method vs random Fourier features: a theoretical and empirical comparison
  publication-title: Adv. Neural Inf. Process. Syst.
– volume: 36
  start-page: A1500
  issue: 4
  year: 2014
  ident: 10.1016/j.acha.2022.08.003_br0140
  article-title: Active subspace methods in theory and practice: applications to Kriging surfaces
  publication-title: SIAM J. Sci. Comput.
  doi: 10.1137/130916138
– year: 2013
  ident: 10.1016/j.acha.2022.08.003_br0230
– ident: 10.1016/j.acha.2022.08.003_br0500
– ident: 10.1016/j.acha.2022.08.003_br0310
– ident: 10.1016/j.acha.2022.08.003_br0490
– volume: 52
  start-page: 1289
  issue: 4
  year: 2006
  ident: 10.1016/j.acha.2022.08.003_br0180
  article-title: Compressed sensing
  publication-title: IEEE Trans. Inf. Theory
  doi: 10.1109/TIT.2006.871582
– start-page: 322
  year: 2019
  ident: 10.1016/j.acha.2022.08.003_br0030
  article-title: Fine-grained analysis of optimization and generalization for overparameterized two-layer neural networks
– volume: 326
  start-page: 402
  year: 2017
  ident: 10.1016/j.acha.2022.08.003_br0150
  article-title: A near-stationary subspace for ridge approximation
  publication-title: Comput. Methods Appl. Mech. Eng.
  doi: 10.1016/j.cma.2017.07.038
– start-page: 555
  year: 2008
  ident: 10.1016/j.acha.2022.08.003_br0400
  article-title: Uniform approximation of functions with random bases
– start-page: 6194
  year: 2020
  ident: 10.1016/j.acha.2022.08.003_br0620
  article-title: Sparse shrunk additive models
– start-page: 7322
  year: 2018
  ident: 10.1016/j.acha.2022.08.003_br0660
  article-title: Streaming kernel PCA with O(n) random
– volume: 302
  start-page: 1
  year: 2015
  ident: 10.1016/j.acha.2022.08.003_br0160
  article-title: Exploiting active subspaces to quantify uncertainty in the numerical simulation of the HyShot II scramjet
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2015.09.001
– start-page: 1009
  year: 2009
  ident: 10.1016/j.acha.2022.08.003_br0640
  article-title: Sparse additive models
  publication-title: J. R. Stat. Soc., Ser. B, Stat. Methodol.
  doi: 10.1111/j.1467-9868.2009.00718.x
– start-page: 2456
  year: 2014
  ident: 10.1016/j.acha.2022.08.003_br0530
  article-title: Sparse random feature algorithm as coordinate descent in Hilbert space
  publication-title: Adv. Neural Inf. Process. Syst.
– ident: 10.1016/j.acha.2022.08.003_br0270
– volume: 39
  start-page: 930
  issue: 3
  year: 1993
  ident: 10.1016/j.acha.2022.08.003_br0040
  article-title: Universal approximation bounds for superpositions of a sigmoidal function
  publication-title: IEEE Trans. Inf. Theory
  doi: 10.1109/18.256500
– ident: 10.1016/j.acha.2022.08.003_br0200
– start-page: 2554
  year: 2014
  ident: 10.1016/j.acha.2022.08.003_br0340
  article-title: UQlab: a framework for uncertainty quantification in Matlab
– start-page: 3905
  year: 2019
  ident: 10.1016/j.acha.2022.08.003_br0320
  article-title: Towards a unified analysis of random Fourier features
– start-page: 567
  year: 2012
  ident: 10.1016/j.acha.2022.08.003_br0550
  article-title: Bayesian lasso for semiparametric structural equation models
  publication-title: Biometrics
  doi: 10.1111/j.1541-0420.2012.01751.x
– volume: 18
  start-page: 661
  issue: 3
  year: 2018
  ident: 10.1016/j.acha.2022.08.003_br0020
  article-title: Infinite-dimensional compressed sensing and function interpolation
  publication-title: Found. Comput. Math.
  doi: 10.1007/s10208-017-9350-3
– volume: 52
  start-page: 5406
  issue: 12
  year: 2006
  ident: 10.1016/j.acha.2022.08.003_br0090
  article-title: Near-optimal signal recovery from random projections: universal encoding strategies?
  publication-title: IEEE Trans. Inf. Theory
  doi: 10.1109/TIT.2006.885507
– ident: 10.1016/j.acha.2022.08.003_br0210
– start-page: 161
  year: 2012
  ident: 10.1016/j.acha.2022.08.003_br0120
  article-title: Efficient kernel clustering using random Fourier features
– volume: 164
  start-page: 517
  issue: 5
  year: 2012
  ident: 10.1016/j.acha.2022.08.003_br0430
  article-title: Sparse Legendre expansions via l1-minimization
  publication-title: J. Approx. Theory
  doi: 10.1016/j.jat.2012.01.008
– year: 2008
  ident: 10.1016/j.acha.2022.08.003_br0450
– start-page: 172
  year: 1927
  ident: 10.1016/j.acha.2022.08.003_br0250
  article-title: About the descriptive content of quantum theoretical kinematics and mechanics
  publication-title: Z. Phys.
  doi: 10.1007/BF01397280
– volume: 33
  start-page: 125
  issue: 1
  year: 2011
  ident: 10.1016/j.acha.2022.08.003_br0170
  article-title: Approximation of functions of few variables in high dimensions
  publication-title: Constr. Approx.
  doi: 10.1007/s00365-010-9105-8
– start-page: 890
  year: 2009
  ident: 10.1016/j.acha.2022.08.003_br0680
  article-title: Probing the Pareto frontier for basis pursuit solutions
  publication-title: SIAM J. Sci. Comput.
  doi: 10.1137/080714488
– volume: 55
  start-page: 3388
  issue: 7
  year: 2009
  ident: 10.1016/j.acha.2022.08.003_br0070
  article-title: On recovery of sparse signals via ℓ1 minimization
  publication-title: IEEE Trans. Inf. Theory
  doi: 10.1109/TIT.2009.2021377
– ident: 10.1016/j.acha.2022.08.003_br0240
– ident: 10.1016/j.acha.2022.08.003_br0390
– volume: 230
  start-page: 3015
  issue: 8
  year: 2011
  ident: 10.1016/j.acha.2022.08.003_br0190
  article-title: A non-adapted sparse approximation of PDEs with stochastic inputs
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2011.01.002
– volume: 34
  start-page: 123
  issue: 1
  year: 1962
  ident: 10.1016/j.acha.2022.08.003_br0060
  article-title: The perceptron: a model for brain functioning. I
  publication-title: Rev. Mod. Phys.
  doi: 10.1103/RevModPhys.34.123
– ident: 10.1016/j.acha.2022.08.003_br0380
– start-page: 43
  year: 2013
  ident: 10.1016/j.acha.2022.08.003_br0580
  article-title: Learning with infinitely many features
– year: 2006
  ident: 10.1016/j.acha.2022.08.003_br0360
  article-title: Randomized clustering forests for building fast and discriminative visual vocabularies
– start-page: 1313
  year: 2008
  ident: 10.1016/j.acha.2022.08.003_br0410
  article-title: Weighted sums of random kitchen sinks: replacing minimization with randomization in learning
  publication-title: Adv. Neural Inf. Process. Syst.
– volume: 59
  start-page: 1207
  issue: 8
  year: 2006
  ident: 10.1016/j.acha.2022.08.003_br0080
  article-title: Stable signal recovery from incomplete and inaccurate measurements
  publication-title: Commun. Pure Appl. Math.
  doi: 10.1002/cpa.20124
– volume: 311
  start-page: 1415
  issue: 87
  year: 2018
  ident: 10.1016/j.acha.2022.08.003_br0130
  article-title: Polynomial approximation via compressed sensing of high-dimensional functions on lower sets
  publication-title: Math. Comput.
– start-page: 827
  year: 2019
  ident: 10.1016/j.acha.2022.08.003_br0670
  article-title: On kernel derivative approximation with random Fourier features
– start-page: 93
  year: 2017
  ident: 10.1016/j.acha.2022.08.003_br0010
  article-title: Compressed sensing approaches for polynomial approximation of high-dimensional functions
– start-page: 1
  year: 2007
  ident: 10.1016/j.acha.2022.08.003_br0540
  article-title: Detecting disease-causing genes by LASSO-Patternsearch algorithm
  publication-title: BMC Proc.
– volume: 18
  start-page: 1435
  issue: 4
  year: 2020
  ident: 10.1016/j.acha.2022.08.003_br0470
  article-title: Extracting structured dynamical systems using sparse optimization with very few samples
  publication-title: Multiscale Model. Simul.
  doi: 10.1137/18M1194730
– start-page: 544
  year: 2007
  ident: 10.1016/j.acha.2022.08.003_br0570
  article-title: l-1 regularization in infinite dimensional feature spaces
– start-page: 35
  year: 2017
  ident: 10.1016/j.acha.2022.08.003_br0650
  article-title: Generalization properties of learning with random features
– volume: 79
  start-page: 953
  issue: 270
  year: 2010
  ident: 10.1016/j.acha.2022.08.003_br0290
  article-title: On decompositions of multivariate functions
  publication-title: Math. Comput.
  doi: 10.1090/S0025-5718-09-02319-9
– start-page: 197
  year: 2017
  ident: 10.1016/j.acha.2022.08.003_br0630
  article-title: Group sparse additive machine
– volume: 78
  start-page: 3279
  issue: 6
  year: 2018
  ident: 10.1016/j.acha.2022.08.003_br0460
  article-title: Extracting sparse high-dimensional dynamics from limited data
  publication-title: SIAM J. Appl. Math.
  doi: 10.1137/18M116798X
– volume: 12
  start-page: 229
  issue: 2
  year: 2012
  ident: 10.1016/j.acha.2022.08.003_br0220
  article-title: Learning functions of few arbitrary linear parameters in high dimensions
  publication-title: Found. Comput. Math.
  doi: 10.1007/s10208-012-9115-y
– volume: 69
  start-page: 593
  issue: 4
  year: 2004
  ident: 10.1016/j.acha.2022.08.003_br0330
  article-title: On the computational power of circuits of spiking neurons
  publication-title: J. Comput. Syst. Sci.
  doi: 10.1016/j.jcss.2004.04.001
– start-page: 627
  year: 2015
  ident: 10.1016/j.acha.2022.08.003_br0560
  article-title: Learning interactions via hierarchical group-lasso regularization
  publication-title: J. Comput. Graph. Stat.
  doi: 10.1080/10618600.2014.938812
– ident: 10.1016/j.acha.2022.08.003_br0590
– volume: 20
  start-page: 608
  issue: 1
  year: 1992
  ident: 10.1016/j.acha.2022.08.003_br0280
  article-title: A simple lemma on greedy approximation in Hilbert space and convergence rates for projection pursuit regression and neural network training
  publication-title: Ann. Stat.
  doi: 10.1214/aos/1176348546
SSID ssj0011459
Score 2.5503297
Snippet Random feature methods have been successful in various machine learning tasks, are easy to compute, and come with theoretical accuracy bounds. They serve as an...
SourceID crossref
elsevier
SourceType Enrichment Source
Index Database
Publisher
StartPage 310
SubjectTerms Compressive sensing
Generalization error
Random features
Sparse optimization
Title Generalization bounds for sparse random feature expansions
URI https://dx.doi.org/10.1016/j.acha.2022.08.003
Volume 62
WOSCitedRecordID wos000870987200003&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  customDbUrl:
  eissn: 1096-603X
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0011459
  issn: 1063-5203
  databaseCode: AIEXJ
  dateStart: 20211211
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1Li9RAEG7GWQ96EJ-4vsjBWwhkupN0t7dFVlRwEWYX5hb6ia5jZpjHsnryp1v9SGdmlUUFL2HSTCdN1Ud1daXqK4Re8rJUtuSi0EqbohINLriltmjqmgnDDMay8s0m6MkJm834x9HoR18LczGnXccuL_nyv6oaxkDZrnT2L9SdHgoD8BuUDldQO1z_SPGRSDrWV-bS9U3ypAs5GI_V2uSwO-nF19waz-npOP5hu0phu56RNnqnsextud30UUPHde3b5ojIZzIYsbVjH_DWRi7mVnxP2RtTl5vvOrH4nU5800MB2tR3FY4Z3kPMe6m2bvjMbr_sBiYwuRKYSBUzewmdcAIlcPgtg1EzcYw3RVP6vsDJKkcbHcwqiamvYYcm4UvOL8Y_xCHOAYifHKMUxp6ctX_VHqn21K3DLQP7UqWa3EAHmNacjdHB0bvj2fv0JWpS-YZ7ad2x8CrkCF590--dmx2H5fQuuhNPGtlRQMg9NDLdfXR7h38S7j4k0t71A_RqHzlZQE4GyMkCcrKAnCwiJxuQ8xCdvTk-ff22iK01CkWqalNoWhksueTCEq6kJeBpS6xrl1FMDG5oo5XV3JUq25JIpRUF315TJYmYCJDXIzTuFp15jDImYRYT2mBaVkoxIZqGMUFLMOjMKHuIJr1MWhV55137k3nbJxiet06OrZNj63qiluQQ5WnOMrCuXPvvuhd1G_3G4A-2gIxr5j35x3lP0a0B78_QeLPamufoprrYfF6vXkQA_QTpgZPW
linkProvider Elsevier
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Generalization+bounds+for+sparse+random+feature+expansions&rft.jtitle=Applied+and+computational+harmonic+analysis&rft.au=Hashemi%2C+Abolfazl&rft.au=Schaeffer%2C+Hayden&rft.au=Shi%2C+Robert&rft.au=Topcu%2C+Ufuk&rft.date=2023-01-01&rft.pub=Elsevier+Inc&rft.issn=1063-5203&rft.eissn=1096-603X&rft.volume=62&rft.spage=310&rft.epage=330&rft_id=info:doi/10.1016%2Fj.acha.2022.08.003&rft.externalDocID=S1063520322000653
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1063-5203&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1063-5203&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1063-5203&client=summon