Approximate Softmax Functions for Energy-Efficient Deep Neural Networks

Approximate computing has emerged as a new paradigm that provides power-efficient and high-performance arithmetic designs by relaxing the stringent requirement of accuracy. Nonlinear functions (such as softmax , rectified linear unit ( ReLU ), Tanh , and Sigmoid ) are extensively used in deep neural...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on very large scale integration (VLSI) systems Ročník 31; číslo 1; s. 1 - 13
Hlavní autoři: Chen, Ke, Gao, Yue, Waris, Haroon, Liu, Weiqiang, Lombardi, Fabrizio
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 01.01.2023
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
ISSN:1063-8210, 1557-9999
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Approximate computing has emerged as a new paradigm that provides power-efficient and high-performance arithmetic designs by relaxing the stringent requirement of accuracy. Nonlinear functions (such as softmax , rectified linear unit ( ReLU ), Tanh , and Sigmoid ) are extensively used in deep neural networks (DNNs). However, they incur significant power dissipation due to the high circuit complexity. As DNNs are error-tolerant, the design of approximation-linear functions is possible and desired. In this article, the design of an approximate softmax function (AxSF) is proposed. AxSF is based on a double hybrid structure (DHS). AxSF divides the input of the softmax function into two parts for different processing methods. The most significant bits (MSBs) are processed with lookup tables (LUTs) and an exact restoring array divider (EXDr). Taylor's expansion and a logarithmic divider are used for the less significant bits (LSBs). An improved DHS (IDHS) is also proposed to reduce the hardware complexity. In IDHS, a novel Booth multiplier is utilized for the hybrid scheme to improve the partial product generation and compression, while the truncated implementation is applied to the divider unit. The proposed DHS and IDHS are compared with existing softmax designs. The results show that the proposed approximate softmax design reduces hardware by 48% and delay by 54% while retaining a high accuracy.
AbstractList Approximate computing has emerged as a new paradigm that provides power-efficient and high-performance arithmetic designs by relaxing the stringent requirement of accuracy. Nonlinear functions (such as softmax, rectified linear unit (ReLU), Tanh, and Sigmoid) are extensively used in deep neural networks (DNNs). However, they incur significant power dissipation due to the high circuit complexity. As DNNs are error-tolerant, the design of approximation-linear functions is possible and desired. In this article, the design of an approximate softmax function (AxSF) is proposed. AxSF is based on a double hybrid structure (DHS). AxSF divides the input of the softmax function into two parts for different processing methods. The most significant bits (MSBs) are processed with lookup tables (LUTs) and an exact restoring array divider (EXDr). Taylor’s expansion and a logarithmic divider are used for the less significant bits (LSBs). An improved DHS (IDHS) is also proposed to reduce the hardware complexity. In IDHS, a novel Booth multiplier is utilized for the hybrid scheme to improve the partial product generation and compression, while the truncated implementation is applied to the divider unit. The proposed DHS and IDHS are compared with existing softmax designs. The results show that the proposed approximate softmax design reduces hardware by 48% and delay by 54% while retaining a high accuracy.
Author Waris, Haroon
Liu, Weiqiang
Gao, Yue
Chen, Ke
Lombardi, Fabrizio
Author_xml – sequence: 1
  givenname: Ke
  orcidid: 0000-0003-0981-3166
  surname: Chen
  fullname: Chen, Ke
  organization: College of Electronic Information and Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China
– sequence: 2
  givenname: Yue
  surname: Gao
  fullname: Gao, Yue
  organization: College of Electronic Information and Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China
– sequence: 3
  givenname: Haroon
  orcidid: 0000-0003-4670-3919
  surname: Waris
  fullname: Waris, Haroon
  organization: College of Electronic Information and Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China
– sequence: 4
  givenname: Weiqiang
  orcidid: 0000-0001-8398-8648
  surname: Liu
  fullname: Liu, Weiqiang
  organization: College of Electronic Information and Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China
– sequence: 5
  givenname: Fabrizio
  orcidid: 0000-0003-3152-3245
  surname: Lombardi
  fullname: Lombardi, Fabrizio
  organization: Department of Electrical and Computer Engineering, Northeastern University, Boston, MA, USA
BookMark eNp9kMFOAjEQhhuDiYi-gF428bzYdtvt9kgQkIToAfTalDo1i7Bd2xLh7S1CPHhwDjNzmH_-me8SdRrXAEI3BPcJwfJ-8TqbT_sUU9ovKGWYkDPUJZyLXKbopB6XRV5Rgi_QZQgrjAljEnfRZNC23u3qjY6QzZ2NG73LxtvGxNo1IbPOZ6MG_Ps-H1lbmxqamD0AtNkTbL1epxK_nP8IV-jc6nWA61PtoZfxaDF8zGfPk-lwMMsNlTzmDEOFNSec2JSWS7DAmLVQ8iXXBoBrXXBNSlMIYTARnPG30nJrWCnKyhRFD90d96arP7cQolq5rW-SpaKCCyKk4DRN0eOU8S4ED1a1Pr3o94pgdQCmfoCpAzB1ApZE1R-RqaM-cIhe1-v_pbdHaQ0Av15SlhUhsvgGLYR7Gw
CODEN ITCOB4
CitedBy_id crossref_primary_10_1109_TCSII_2025_3594105
crossref_primary_10_1016_j_conengprac_2025_106275
crossref_primary_10_1109_TCSI_2024_3406167
crossref_primary_10_1109_TCSI_2025_3559069
crossref_primary_10_1109_TCSI_2024_3392807
crossref_primary_10_1080_19392699_2025_2565800
crossref_primary_10_1109_TC_2024_3398512
crossref_primary_10_1007_s00034_025_03267_7
crossref_primary_10_1109_LSENS_2024_3497148
crossref_primary_10_1088_1361_6501_ad076c
crossref_primary_10_1109_TVLSI_2025_3553069
crossref_primary_10_1007_s10470_025_02490_1
crossref_primary_10_1007_s11227_024_06212_8
crossref_primary_10_1142_S0218001425580030
crossref_primary_10_1109_ACCESS_2025_3592609
crossref_primary_10_1109_TCSI_2024_3443270
Cites_doi 10.1145/1978542.1978559
10.23919/DATE.2017.7927254
10.1109/TC.2012.146
10.1109/ICSICT.2018.8565706
10.1109/DATE.2011.5763154
10.1038/323533a0
10.1109/TEC.1958.5222579
10.1109/CVPR.2015.7298594
10.1109/VDAT50263.2020.9190498
10.1109/TCSI.2015.2403036
10.1162/neco.1997.9.8.1735
10.1109/TVLSI.2020.3015391
10.1007/s11263-009-0275-4
10.1145/2830772.2830810
10.1109/JPROC.2020.3006451
10.1109/TEC.1962.5219391
10.1109/ACCESS.2020.2985345
10.1109/ICAIIS49377.2020.9194894
10.1109/TC.2015.2494005
10.1109/TVLSI.2019.2905242
10.1109/TC.2017.2672976
10.1109/DATE.2008.4484850
10.1109/IWOFC48002.2019.9078446
10.1109/SOCC.2016.7905501
10.1109/TCSI.2021.3069168
10.1109/ICCAD.2013.6691096
10.1007/978-3-030-20870-7_7
10.1162/neco.2006.18.7.1527
10.1109/MLSP.2008.4685487
10.1109/ISCAS45731.2020.9180870
10.1145/2228360.2228509
10.1145/3299874.3317988
10.1109/TPAMI.2016.2577031
10.1109/TCSI.2018.2792902
10.1007/s00521-007-0086-x
10.1145/3020078.3021745
10.1088/1742-6596/368/1/012030
10.1109/OJCS.2021.3051643
10.1038/nature12346
10.1109/ASAP49362.2020.00017
10.1145/3079856.3080246
10.1109/TCAD.2018.2871198
10.1109/MSP.2012.2205597
10.1109/TCSII.2020.3002564
10.1109/TETC.2021.3109127
10.1109/ETS.2013.6569370
10.1109/DATE.2012.6176685
10.1145/3174243.3174253
10.1007/978-1-4471-0487-2
10.1145/2742060.2743759
10.1109/ISLPED.2019.8824959
10.1109/TVLSI.2018.2815603
10.1109/MICRO.2007.9
10.1109/TSUSC.2020.3004980
10.1109/CVPR.2014.222
10.1109/ICASID.2018.8693206
10.1145/2463209.2488873
ContentType Journal Article
Copyright Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023
Copyright_xml – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023
DBID 97E
RIA
RIE
AAYXX
CITATION
7SP
8FD
L7M
DOI 10.1109/TVLSI.2022.3224011
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
Electronics & Communications Abstracts
Technology Research Database
Advanced Technologies Database with Aerospace
DatabaseTitle CrossRef
Technology Research Database
Advanced Technologies Database with Aerospace
Electronics & Communications Abstracts
DatabaseTitleList Technology Research Database

Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1557-9999
EndPage 13
ExternalDocumentID 10_1109_TVLSI_2022_3224011
9968119
Genre orig-research
GrantInformation_xml – fundername: National Natural Science Foundation of China
  grantid: 62101252; 62022041; 61871216
  funderid: 10.13039/501100001809
– fundername: NSF through the following grants: “An Integrated Framework for System-Level Approximate Computing”, and “Neural-Network-based Stochastic Computing Architectures With Applications to Machine Learning”
GroupedDBID -~X
.DC
0R~
29I
4.4
5GY
6IK
97E
AAJGR
AASAJ
AAWTH
ABQJQ
ABVLG
ACGFS
ACIWK
AENEX
AGQYO
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
HZ~
IEDLZ
IFIPE
IPLJI
JAVBF
LAI
M43
O9-
OCL
P2P
RIA
RIE
RNS
TN5
3EH
5VS
AAYXX
ABFSI
AETIX
AGSQL
AI.
AIBXA
ALLEH
CITATION
E.L
EJD
H~9
ICLAB
IFJZH
VH1
7SP
8FD
L7M
ID FETCH-LOGICAL-c295t-40e80a5151f515bbefe44ffe65b5acee5aa35a16c377c017545d6f5fc46768c33
IEDL.DBID RIE
ISICitedReferencesCount 23
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000911286400002&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1063-8210
IngestDate Sun Nov 30 04:09:14 EST 2025
Sat Nov 29 03:36:20 EST 2025
Tue Nov 18 21:42:11 EST 2025
Tue Nov 25 14:44:28 EST 2025
IsPeerReviewed true
IsScholarly true
Issue 1
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c295t-40e80a5151f515bbefe44ffe65b5acee5aa35a16c377c017545d6f5fc46768c33
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ORCID 0000-0003-0981-3166
0000-0001-8398-8648
0000-0003-3152-3245
0000-0003-4670-3919
PQID 2757179752
PQPubID 85424
PageCount 13
ParticipantIDs crossref_primary_10_1109_TVLSI_2022_3224011
ieee_primary_9968119
proquest_journals_2757179752
crossref_citationtrail_10_1109_TVLSI_2022_3224011
PublicationCentury 2000
PublicationDate 2023-01-01
PublicationDateYYYYMMDD 2023-01-01
PublicationDate_xml – month: 01
  year: 2023
  text: 2023-01-01
  day: 01
PublicationDecade 2020
PublicationPlace New York
PublicationPlace_xml – name: New York
PublicationTitle IEEE transactions on very large scale integration (VLSI) systems
PublicationTitleAbbrev TVLSI
PublicationYear 2023
Publisher IEEE
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml – name: IEEE
– name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References ref57
ref12
ref56
ref15
ref59
ref14
ref53
ref52
ref55
ref10
ref54
ref19
ref18
ref51
ref50
Lisboa (ref17) 2000
ref46
ref45
ref48
Dai (ref58) 2016
ref47
ref42
ref41
ref44
ref43
ref49
ref8
ref7
ref9
ref4
ref3
ref6
ref5
ref40
Collobert (ref11) 2011; 12
ref35
ref34
Mcnelis (ref16) 2004
ref37
ref36
ref31
ref30
ref33
ref32
ref2
Quinonerocandela (ref13) 2009
ref1
ref39
ref38
ref24
ref23
ref26
ref25
ref20
ref22
ref21
ref28
ref27
ref29
ref60
ref61
References_xml – ident: ref1
  doi: 10.1145/1978542.1978559
– ident: ref46
  doi: 10.23919/DATE.2017.7927254
– ident: ref54
  doi: 10.1109/TC.2012.146
– ident: ref23
  doi: 10.1109/ICSICT.2018.8565706
– ident: ref27
  doi: 10.1109/DATE.2011.5763154
– ident: ref8
  doi: 10.1038/323533a0
– ident: ref45
  doi: 10.1109/TEC.1958.5222579
– ident: ref10
  doi: 10.1109/CVPR.2015.7298594
– ident: ref52
  doi: 10.1109/VDAT50263.2020.9190498
– ident: ref25
  doi: 10.1109/TCSI.2015.2403036
– ident: ref56
  doi: 10.1162/neco.1997.9.8.1735
– ident: ref55
  doi: 10.1109/TVLSI.2020.3015391
– ident: ref60
  doi: 10.1007/s11263-009-0275-4
– ident: ref19
  doi: 10.1145/2830772.2830810
– ident: ref33
  doi: 10.1109/JPROC.2020.3006451
– ident: ref34
  doi: 10.1109/TEC.1962.5219391
– ident: ref38
  doi: 10.1109/ACCESS.2020.2985345
– ident: ref53
  doi: 10.1109/ICAIIS49377.2020.9194894
– ident: ref44
  doi: 10.1109/TC.2015.2494005
– ident: ref59
  doi: 10.1109/TVLSI.2019.2905242
– ident: ref32
  doi: 10.1109/TC.2017.2672976
– ident: ref26
  doi: 10.1109/DATE.2008.4484850
– ident: ref41
  doi: 10.1109/IWOFC48002.2019.9078446
– ident: ref47
  doi: 10.1109/SOCC.2016.7905501
– ident: ref39
  doi: 10.1109/TCSI.2021.3069168
– ident: ref30
  doi: 10.1109/ICCAD.2013.6691096
– ident: ref49
  doi: 10.1007/978-3-030-20870-7_7
– year: 2016
  ident: ref58
  article-title: R-FCN: Object detection via region-based fully convolutional networks
  publication-title: arXiv:1605.06409
– ident: ref7
  doi: 10.1162/neco.2006.18.7.1527
– volume-title: Neural Networks in Finance: Gaining Predictive Edge in the Market
  year: 2004
  ident: ref16
– ident: ref18
  doi: 10.1109/MLSP.2008.4685487
– ident: ref40
  doi: 10.1109/ISCAS45731.2020.9180870
– ident: ref28
  doi: 10.1145/2228360.2228509
– ident: ref48
  doi: 10.1145/3299874.3317988
– ident: ref57
  doi: 10.1109/TPAMI.2016.2577031
– ident: ref36
  doi: 10.1109/TCSI.2018.2792902
– ident: ref20
  doi: 10.1007/s00521-007-0086-x
– ident: ref21
  doi: 10.1145/3020078.3021745
– ident: ref14
  doi: 10.1088/1742-6596/368/1/012030
– volume: 12
  start-page: 2493
  year: 2011
  ident: ref11
  article-title: Natural language processing (almost) from scratch
  publication-title: J. Mach. Learn. Res.
– ident: ref61
  doi: 10.1109/OJCS.2021.3051643
– ident: ref15
  doi: 10.1038/nature12346
– ident: ref50
  doi: 10.1109/ASAP49362.2020.00017
– ident: ref31
  doi: 10.1145/3079856.3080246
– ident: ref42
  doi: 10.1109/TCAD.2018.2871198
– ident: ref12
  doi: 10.1109/MSP.2012.2205597
– ident: ref51
  doi: 10.1109/TCSII.2020.3002564
– ident: ref37
  doi: 10.1109/TETC.2021.3109127
– ident: ref4
  doi: 10.1109/ETS.2013.6569370
– ident: ref29
  doi: 10.1109/DATE.2012.6176685
– ident: ref22
  doi: 10.1145/3174243.3174253
– volume-title: Artificial Neural Networks in Biomedicine
  year: 2000
  ident: ref17
  doi: 10.1007/978-1-4471-0487-2
– volume-title: Machine Learning Challenges
  year: 2009
  ident: ref13
– ident: ref6
  doi: 10.1145/2742060.2743759
– ident: ref43
  doi: 10.1109/ISLPED.2019.8824959
– ident: ref2
  doi: 10.1109/TVLSI.2018.2815603
– ident: ref5
  doi: 10.1109/MICRO.2007.9
– ident: ref35
  doi: 10.1109/TSUSC.2020.3004980
– ident: ref9
  doi: 10.1109/CVPR.2014.222
– ident: ref24
  doi: 10.1109/ICASID.2018.8693206
– ident: ref3
  doi: 10.1145/2463209.2488873
SSID ssj0014490
Score 2.4889083
Snippet Approximate computing has emerged as a new paradigm that provides power-efficient and high-performance arithmetic designs by relaxing the stringent requirement...
SourceID proquest
crossref
ieee
SourceType Aggregation Database
Enrichment Source
Index Database
Publisher
StartPage 1
SubjectTerms Accuracy
Adders
Approximate computing
Artificial neural networks
Circuit design
Complexity
Deep neural network (DNN)
Dividers
Energy dissipation
field-programmable gate array (FPGA) implementation
Hardware
Hybrid structures
Linear functions
Logic gates
Lookup tables
Mathematical analysis
Neural networks
nonlinear function
Power management
softmax layer
Table lookup
Taylor series
Title Approximate Softmax Functions for Energy-Efficient Deep Neural Networks
URI https://ieeexplore.ieee.org/document/9968119
https://www.proquest.com/docview/2757179752
Volume 31
WOSCitedRecordID wos000911286400002&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: PRVIEE
  databaseName: IEEE Electronic Library (IEL)
  customDbUrl:
  eissn: 1557-9999
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0014490
  issn: 1063-8210
  databaseCode: RIE
  dateStart: 19930101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1bS8MwFA5z-KAP3qY4ndIH37Rb0-bSPg7dVBhD2JS9lTQ9AUG3sYvs53uSdkNRBF9KH5K0fLmcc5Kc7yPkyghNqQiZz6kwPlMs8DMea18AMK2t4jY41ZKe7Pfj0Sh5qpCbTS4MALjLZ9C0r-4sP5_opd0qa6FvHlPL8bklpSxytTYnBowlBfOAiPwY45h1gkyQtIYvvcEjhoJh2IysBaP0mxFyqio_lmJnX7r7__uzA7JX-pFeu-j4Q1KB8RHZ_cIuWCP3bcsXvnpFnxS8AS6372rlddGOuaHmobfqdVzmn99xPBL4Be8OYOpZwg5su1_cEJ8fk-duZ3j74Je6Cb4OE77AkBDiQKGjQg0-sgwMMGYMCJ5xhUaRKxVxRYWOpNQ4I9GJyoXhRuOiKWIdRSekOp6M4ZR4VMlcqNBQFRsGWaByYGAZ80HSJFdJndA1kKkuScWttsVb6oKLIEkd-KkFPy3Br5PrTZ1pQanxZ-mahXtTskS6Thrr_krLWTdPQ8kxOk0kD89-r3VOdqxcfLGF0iDVxWwJF2Rbfyxe57NLN6A-AXYIyFI
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1bS8MwFD6ICuqDtynOax9802rTJmn7KLq54RyCU_ZW0vQEBJ3iNtnP9yTthqIIvpQ-JGn5cjnnJDnfB3BspGZMhtwXTBqfKx74uUi0LxG51lZxG51qSSfudpN-P72bg9NZLgwiustneGZf3Vl-8arHdqvsnHzzhFmOzwXBecjKbK3ZmQHnack9ICM_oUhmmiITpOe9x859m4LBMDyLrA1j7JsZcroqPxZjZ2Gaa__7t3VYrTxJ76Ls-g2Yw8EmrHzhF6zB9YVlDJ88kVeK3j0tuC9q4jXJkrnB5pG_6jVc7p_fcEwS9AXvCvHNs5Qd1Ha3vCM-3IKHZqN32fIr5QRfh6kYUVCISaDIVWGGHnmOBjk3BqXIhSKzKJSKhGJSR3GsaU6SG1VII4ymZVMmOoq2YX7wOsAd8JiKC6lCw1RiOOaBKpCj5czHmKWFSuvApkBmuqIVt-oWz5kLL4I0c-BnFvysAr8OJ7M6byWpxp-laxbuWckK6TrsT_srq-bdMAtjQfFpGotw9_daR7DU6t12sk67e7MHy1Y8vtxQ2Yf50fsYD2BRf4yehu-HbnB9AgwMy5k
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=Approximate+Softmax+Functions+for+Energy-Efficient+Deep+Neural+Networks&rft.jtitle=IEEE+transactions+on+very+large+scale+integration+%28VLSI%29+systems&rft.au=Chen%2C+Ke&rft.au=Gao%2C+Yue&rft.au=Waris%2C+Haroon&rft.au=Liu%2C+Weiqiang&rft.date=2023-01-01&rft.pub=The+Institute+of+Electrical+and+Electronics+Engineers%2C+Inc.+%28IEEE%29&rft.issn=1063-8210&rft.eissn=1557-9999&rft.volume=31&rft.issue=1&rft.spage=4&rft_id=info:doi/10.1109%2FTVLSI.2022.3224011&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1063-8210&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1063-8210&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1063-8210&client=summon