Rescuing memristor-based neuromorphic design with high defects
Memristor-based synaptic network has been widely investigated and applied to neuromorphic computing systems for the fast computation and low design cost. As memristors continue to mature and achieve higher density, bit failures within crossbar arrays can become a critical issue. These can degrade th...
Gespeichert in:
| Veröffentlicht in: | 2017 54th ACM/EDAC/IEEE Design Automation Conference (DAC) S. 1 - 6 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Tagungsbericht |
| Sprache: | Englisch |
| Veröffentlicht: |
IEEE
01.06.2017
|
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Abstract | Memristor-based synaptic network has been widely investigated and applied to neuromorphic computing systems for the fast computation and low design cost. As memristors continue to mature and achieve higher density, bit failures within crossbar arrays can become a critical issue. These can degrade the computation accuracy significantly. In this work, we propose a defect rescuing design to restore the computation accuracy. In our proposed design, significant weights in a specified network are first identified and retraining and remapping algorithms are described. For a two layer neural network with 92.64% classification accuracy on MNIST digit recognition, our evaluation based on real device testing shows that our design can recover almost its full performance when 20% random defects are present. |
|---|---|
| AbstractList | Memristor-based synaptic network has been widely investigated and applied to neuromorphic computing systems for the fast computation and low design cost. As memristors continue to mature and achieve higher density, bit failures within crossbar arrays can become a critical issue. These can degrade the computation accuracy significantly. In this work, we propose a defect rescuing design to restore the computation accuracy. In our proposed design, significant weights in a specified network are first identified and retraining and remapping algorithms are described. For a two layer neural network with 92.64% classification accuracy on MNIST digit recognition, our evaluation based on real device testing shows that our design can recover almost its full performance when 20% random defects are present. |
| Author | Hai Li Strachan, John Paul Miao Hu Chenchen Liu |
| Author_xml | – sequence: 1 surname: Chenchen Liu fullname: Chenchen Liu email: CHL192@pitt.edu organization: Dept. of Electr. & Comput. Eng., Univ. of Pittsburgh, Pittsburgh, PA, USA – sequence: 2 surname: Miao Hu fullname: Miao Hu email: miao.hu@hpe.com organization: Hewlett Packard Labs., Palo Alto, CA, USA – sequence: 3 givenname: John Paul surname: Strachan fullname: Strachan, John Paul email: john-paul.strachan@hpe.com organization: Hewlett Packard Labs., Palo Alto, CA, USA – sequence: 4 surname: Hai Li fullname: Hai Li email: hai.li@duke.edu organization: Dept. of Electr. & Comput. Eng., Duke Univ., Durham, NC, USA |
| BookMark | eNotzE9LwzAYgPEIenDTswcv-QKdb5I2fXMRZPgPBsKY55Emb9qATUfSIX57C3r6wXN4VuwyTYkYuxOwEaJuHhRooZXZLEol4IKtlgqqNrJtr9njnoo7x9TzkcYcyzzlqrOFPE90ztM45dMQHfdUYp_4d5wHPsR-WEIgN5cbdhXsV6Hbf9fs8-X5sH2rdh-v79unXWUl4lwFsFoHwoAtgaKmsy54EoRWBtIBNBlADx4bI6xDsSgD1J3y3ktsrVqz-79vJKLjKcfR5p8jgoa6MeoXp0RGAQ |
| ContentType | Conference Proceeding |
| DBID | 6IE 6IH CBEJK RIE RIO |
| DOI | 10.1145/3061639.3062310 |
| DatabaseName | IEEE Electronic Library (IEL) Conference Proceedings IEEE Proceedings Order Plan (POP) 1998-present by volume IEEE Xplore All Conference Proceedings IEEE Electronic Library (IEL) IEEE Proceedings Order Plans (POP) 1998-present |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| EISBN | 1450349277 9781450349277 |
| EndPage | 6 |
| ExternalDocumentID | 8060459 |
| Genre | orig-research |
| GroupedDBID | 6IE 6IH CBEJK RIE RIO |
| ID | FETCH-LOGICAL-a288t-f0a66fe8f87e03e5bacfde1e8a2fe6f06e908d0d8591ac818592f04b3ddd287a3 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 85 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000424895400175&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| IngestDate | Thu Jun 29 18:36:44 EDT 2023 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | true |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-a288t-f0a66fe8f87e03e5bacfde1e8a2fe6f06e908d0d8591ac818592f04b3ddd287a3 |
| OpenAccessLink | http://dl.acm.org/ft_gateway.cfm?id=3062310&type=pdf |
| PageCount | 6 |
| ParticipantIDs | ieee_primary_8060459 |
| PublicationCentury | 2000 |
| PublicationDate | 2017-June |
| PublicationDateYYYYMMDD | 2017-06-01 |
| PublicationDate_xml | – month: 06 year: 2017 text: 2017-June |
| PublicationDecade | 2010 |
| PublicationTitle | 2017 54th ACM/EDAC/IEEE Design Automation Conference (DAC) |
| PublicationTitleAbbrev | DAC |
| PublicationYear | 2017 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| Score | 2.4656465 |
| Snippet | Memristor-based synaptic network has been widely investigated and applied to neuromorphic computing systems for the fast computation and low design cost. As... |
| SourceID | ieee |
| SourceType | Publisher |
| StartPage | 1 |
| SubjectTerms | Degradation Hardware Memristors Neural networks Neuromorphics Redundancy Training |
| Title | Rescuing memristor-based neuromorphic design with high defects |
| URI | https://ieeexplore.ieee.org/document/8060459 |
| WOSCitedRecordID | wos000424895400175&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 | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV29S0MxED_a4uCk0orfZHA0bV5eXj4WF7E4SCmi0K3kJRfo0FZeW_9-k_RREVycktyQcLnhd7nc_Q7gvpbCRsNyGrhiVCiU1EaYo4UzonAMK-Yyz-yrmkz0bGamHXg41MIgYk4-w2Ga5r98v3a7FCob6cT0UpkudJWS-1qtlq2nENUoOr_RuTDDOCan5Ve7lIwW45P_nXMKg5-yOzI9AMoZdHDVh8c33LhdXJMlLjMRQEMT9HiSqSiX63hRC0d8zsQgKaxKEgVxFOREjQF8jJ_fn15o2_SAWq71lgZmpQyog1bISqxq64LHArXlAWVgEg3TnvnEO2ddglvDAxN16b2Prx9bnkNvtV7hBRAjNHOhEnFjJqxkFpXRsixrbcuaq3AJ_aT7_HPPazFv1b76W3wNxzxBWo5A3EBv2-zwFo7c13axae6yMb4BDi6NvQ |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV29T0MhEL_UaqKTmtb4LYOjtDwejweLi7GpsTaNqUm3hgdH0qGt6Yd_v0CbGhMXJ-AGyHHD7zjufgdwX0lhgmE59bxkVJQoqQkwRzOrRWYZFswmntle2e-r0UgPavCwq4VBxJR8hq04TX_5bm7XMVTWVpHppdB7sF8IwdmmWmvL15OJoh3c3-Be6FYYo9vyq2FKwovO8f9OOoHmT-EdGewg5RRqOGvA4zsu7TqsyRSniQpgQSP4OJLIKKfzcFUTS1zKxSAxsEoiCXEQpFSNJnx0nodPXbpte0ANV2pFPTNSelRelchyLCpjvcMMleEepWcSNVOOucg8Z2wEXM09E1XunAvvH5OfQX02n-E5EC0Us74QYWMmjGQGS61knlfK5BUv_QU0ou7jzw2zxXir9uXf4js47A7feuPeS__1Co54BLgUj7iG-mqxxhs4sF-ryXJxmwzzDZyakQQ |
| 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%3Abook&rft.genre=proceeding&rft.title=2017+54th+ACM%2FEDAC%2FIEEE+Design+Automation+Conference+%28DAC%29&rft.atitle=Rescuing+memristor-based+neuromorphic+design+with+high+defects&rft.au=Chenchen+Liu&rft.au=Miao+Hu&rft.au=Strachan%2C+John+Paul&rft.au=Hai+Li&rft.date=2017-06-01&rft.pub=IEEE&rft.spage=1&rft.epage=6&rft_id=info:doi/10.1145%2F3061639.3062310&rft.externalDocID=8060459 |