MyML: User-Driven Machine Learning
Machine learning (ML) on resource-constrained edge devices is expensive and often requires offloading computation to the cloud, which may compromise the privacy of user data. In contrast, the type of data processed at edge devices is user specific and limited to few inference classes. In this work,...
Uloženo v:
| Vydáno v: | 2021 58th ACM/IEEE Design Automation Conference (DAC) s. 145 - 150 |
|---|---|
| Hlavní autoři: | , , |
| Médium: | Konferenční příspěvek |
| Jazyk: | angličtina |
| Vydáno: |
IEEE
05.12.2021
|
| Témata: | |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | Machine learning (ML) on resource-constrained edge devices is expensive and often requires offloading computation to the cloud, which may compromise the privacy of user data. In contrast, the type of data processed at edge devices is user specific and limited to few inference classes. In this work, we explore the opportunity of building smaller, user-specific machine learning models, rather than utilizing a generic, compute-intensive machine learning model that caters to a diverse range of users. We first present a hardware-friendly, light-weight pruning technique to create user-specific models directly on mobile platforms, while simultaneously executing inferences. The proposed technique leverages compute sharing between pruning and inference, customizes the retraining backward-pass and chooses a pruning granularity for efficient processing on edge. We then propose architectural support to prune user-specific models on a systolic edge ML inference accelerator. We demonstrate that user-specific models provide a speedup of 2.3\times over the generic model on mobile CPUs. |
|---|---|
| AbstractList | Machine learning (ML) on resource-constrained edge devices is expensive and often requires offloading computation to the cloud, which may compromise the privacy of user data. In contrast, the type of data processed at edge devices is user specific and limited to few inference classes. In this work, we explore the opportunity of building smaller, user-specific machine learning models, rather than utilizing a generic, compute-intensive machine learning model that caters to a diverse range of users. We first present a hardware-friendly, light-weight pruning technique to create user-specific models directly on mobile platforms, while simultaneously executing inferences. The proposed technique leverages compute sharing between pruning and inference, customizes the retraining backward-pass and chooses a pruning granularity for efficient processing on edge. We then propose architectural support to prune user-specific models on a systolic edge ML inference accelerator. We demonstrate that user-specific models provide a speedup of 2.3\times over the generic model on mobile CPUs. |
| Author | Das, Reetuparna Bertacco, Valeria Goyal, Vidushi |
| Author_xml | – sequence: 1 givenname: Vidushi surname: Goyal fullname: Goyal, Vidushi email: vidushi@umich.edu organization: University of Michigan,Ann Arbor,USA – sequence: 2 givenname: Valeria surname: Bertacco fullname: Bertacco, Valeria email: valeria@umich.edu organization: University of Michigan,Ann Arbor,USA – sequence: 3 givenname: Reetuparna surname: Das fullname: Das, Reetuparna email: reetudas@umich.edu organization: University of Michigan,Ann Arbor,USA |
| BookMark | eNotj8tKAzEUQCMoqHW-QITB_Uxz87hJ3JWpj8IMbuy6JOmNBjRKRoT-vYLdnLM7cC7ZafksxNgN8B6Au-V6NYDlRvWCC-idtigFP2GNMxYQtZLCKH7OmnnOgSPXVv3xgt1Oh2m8a7cz1W5d8w-VdvLxLRdqR_K15PJ6xc6Sf5-pOXrBtg_3L8NTNz4_bobV2HkB5rsDp7xNaIyLbr-nGKUIMpJNGnxA8gITBCCtrIaYopeghHUhoiXjHCq5YNf_3UxEu6-aP3w97I4n8hcNVj8w |
| ContentType | Conference Proceeding |
| DBID | 6IE 6IH CBEJK RIE RIO |
| DOI | 10.1109/DAC18074.2021.9586320 |
| DatabaseName | IEEE Electronic Library (IEL) Conference Proceedings IEEE Proceedings Order Plan (POP) 1998-present by volume IEEE Xplore All Conference Proceedings IEEE/IET Electronic Library IEEE Proceedings Order Plans (POP) 1998-present |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE/IET Electronic Library url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| EISBN | 9781665432740 1665432748 |
| EndPage | 150 |
| ExternalDocumentID | 9586320 |
| Genre | orig-research |
| GroupedDBID | 6IE 6IH ACM ALMA_UNASSIGNED_HOLDINGS CBEJK RIE RIO |
| ID | FETCH-LOGICAL-a217t-194a8f6779c9ddecc32b3ce8f51ab6ea26f1b1e54851cfca314289bc68e799643 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 3 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000766079700025&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| IngestDate | Wed Aug 27 02:28:29 EDT 2025 |
| IsPeerReviewed | false |
| IsScholarly | true |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-a217t-194a8f6779c9ddecc32b3ce8f51ab6ea26f1b1e54851cfca314289bc68e799643 |
| PageCount | 6 |
| ParticipantIDs | ieee_primary_9586320 |
| PublicationCentury | 2000 |
| PublicationDate | 2021-Dec.-5 |
| PublicationDateYYYYMMDD | 2021-12-05 |
| PublicationDate_xml | – month: 12 year: 2021 text: 2021-Dec.-5 day: 05 |
| PublicationDecade | 2020 |
| PublicationTitle | 2021 58th ACM/IEEE Design Automation Conference (DAC) |
| PublicationTitleAbbrev | DAC |
| PublicationYear | 2021 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| SSID | ssib060584060 |
| Score | 2.1873658 |
| Snippet | Machine learning (ML) on resource-constrained edge devices is expensive and often requires offloading computation to the cloud, which may compromise the... |
| SourceID | ieee |
| SourceType | Publisher |
| StartPage | 145 |
| SubjectTerms | Buildings Collaboration Computational modeling Data privacy Design automation Privacy Transfer learning |
| Title | MyML: User-Driven Machine Learning |
| URI | https://ieeexplore.ieee.org/document/9586320 |
| WOSCitedRecordID | wos000766079700025&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/eLvHCXMwlV1LSwMxEB5q8eBJpRXfLOLRtM3uJtl4k9bioVt6sNJbyWMiXlppt4L_3iRdK4IXbyEQwkyGGb7MzDcAt05rTpWUhHGrSW6oJYW1hjAmrPBgzPJYIPsyEuNxMZvJSQPudr0wiBiLz7ATljGXb5dmE77KupIVPEs9QN8TQmx7tb5tJ2T3fGzq1U06tCe7g4c-DVQvHgSmtFOf_TVEJcaQ4eH_bj-C9k8zXjLZhZljaOCiBTflZzm6T6begMhgFTxWUsaySExqxtTXNkyHj8_9J1KPOyDK44KKUJmrwnEhpJHe6RiTpTozWDhGleaoUu6opughBqPGGZUFsjSpDS9QyECrdQLNxXKBp5BkHgak0mXoVZ4H_pncxZQk1VIrodQZtIJ88_cto8W8Fu387-0LOAgqjEUc7BKa1WqDV7BvPqq39eo6PsMXYFuGwg |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEB5KFfSk0opvF_Fo2mZ3k2y8SWupuFt6aKW3kqf0spXaCv57k3StCF68hUAIMxlm-DIz3wDcWikpFpwjQrVEqcIaZVorRAjTzIExTUOB7EvOhsNsOuWjGtxte2GMMaH4zLT8MuTy9UKt_VdZm5OMJrED6DskTWO86db6th6f33PRqVO16eAOb_ceutiTvTgYGONWdfrXGJUQRfoH_7v_EJo_7XjRaBtojqBmygbcFJ9Ffh9NnAmh3tL7rKgIhZEmqjhTX5sw6T-OuwNUDTxAwiGDFcI8FZmljHHFndtRKollokxmCRaSGhFTiyU2DmQQrKwSiadL41LRzDDuibWOoV4uSnMCUeKAQMxtYpzSU89Ak9qQlMSSS8GEOIWGl2_2tuG0mFWinf29fQ17g3GRz_Kn4fM57Ht1hpIOcgH11XJtLmFXfazm78ur8CRfmr2KCQ |
| 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=2021+58th+ACM%2FIEEE+Design+Automation+Conference+%28DAC%29&rft.atitle=MyML%3A+User-Driven+Machine+Learning&rft.au=Goyal%2C+Vidushi&rft.au=Bertacco%2C+Valeria&rft.au=Das%2C+Reetuparna&rft.date=2021-12-05&rft.pub=IEEE&rft.spage=145&rft.epage=150&rft_id=info:doi/10.1109%2FDAC18074.2021.9586320&rft.externalDocID=9586320 |