Robust Aggregation for Federated Learning
We present a novel approach to federated learning that endows its aggregation process with greater robustness to potential poisoning of local data or model parameters of participating devices. The proposed approach, Robust Federated Aggregation (RFA), relies on the aggregation of updates using the g...
Saved in:
| Published in: | IEEE transactions on signal processing Vol. 70; pp. 1142 - 1154 |
|---|---|
| Main Authors: | , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
New York
IEEE
2022
The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Subjects: | |
| ISSN: | 1053-587X, 1941-0476 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | We present a novel approach to federated learning that endows its aggregation process with greater robustness to potential poisoning of local data or model parameters of participating devices. The proposed approach, Robust Federated Aggregation (RFA), relies on the aggregation of updates using the geometric median, which can be computed efficiently using a Weiszfeld-type algorithm. RFA is agnostic to the level of corruption and aggregates model updates without revealing each device's individual contribution. We establish the convergence of the robust federated learning algorithm for the stochastic learning of additive models with least squares. We also offer two variants of RFA: a faster one with one-step robust aggregation, and another one with on-device personalization. We present experimental results with additive models and deep networks for three tasks in computer vision and natural language processing. The experiments show that RFA is competitive with the classical aggregation when the level of corruption is low, while demonstrating greater robustness under high corruption. |
|---|---|
| AbstractList | We present a novel approach to federated learning that endows its aggregation process with greater robustness to potential poisoning of local data or model parameters of participating devices. The proposed approach, Robust Federated Aggregation (RFA), relies on the aggregation of updates using the geometric median, which can be computed efficiently using a Weiszfeld-type algorithm. RFA is agnostic to the level of corruption and aggregates model updates without revealing each device’s individual contribution. We establish the convergence of the robust federated learning algorithm for the stochastic learning of additive models with least squares. We also offer two variants of RFA: a faster one with one-step robust aggregation, and another one with on-device personalization. We present experimental results with additive models and deep networks for three tasks in computer vision and natural language processing. The experiments show that RFA is competitive with the classical aggregation when the level of corruption is low, while demonstrating greater robustness under high corruption. |
| Author | Pillutla, Krishna Harchaoui, Zaid Kakade, Sham M. |
| Author_xml | – sequence: 1 givenname: Krishna orcidid: 0000-0002-1262-8466 surname: Pillutla fullname: Pillutla, Krishna email: pillutla@cs.washington.edu organization: University of Washington, Seattle, WA, USA – sequence: 2 givenname: Sham M. surname: Kakade fullname: Kakade, Sham M. email: sham@seas.harvard.edu organization: Harvard University, Cambridge, MA, USA – sequence: 3 givenname: Zaid surname: Harchaoui fullname: Harchaoui, Zaid email: zaid@uw.edu organization: University of Washington, Seattle, WA, USA |
| BookMark | eNp9kL9LAzEUx4NUsK3ugsuBk8PVvCR3ScZSrAoFRSu4hTSXHFfqpSbp4H9vaouDg9N7w_fzfnxGaND73iJ0CXgCgOXt8vV5QjAhEwoVBVqdoCFIBiVmvB7kHle0rAR_P0OjGNcYA2OyHqKbF7_axVRM2zbYVqfO94XzoZjbxgadbFMsrA5917fn6NTpTbQXxzpGb_O75eyhXDzdP86mi9IQCakUGDfAhTaidoRZa6nOuzl1TjjCzQo0MVAJJiR3EqxeCSMcI42TzjUcCB2j68PcbfCfOxuTWvtd6PNKRWoqJQVMIKfqQ8oEH2OwTpku_Zyfgu42CrDaa1FZi9prUUctGcR_wG3oPnT4-g-5OiBdfuc3LjkBAEG_AeXzbgM |
| CODEN | ITPRED |
| CitedBy_id | crossref_primary_10_1016_j_conengprac_2025_106366 crossref_primary_10_1016_j_knosys_2023_110813 crossref_primary_10_1109_JIOT_2023_3272334 crossref_primary_10_1145_3616537 crossref_primary_10_1109_TIFS_2025_3576594 crossref_primary_10_1016_j_jisa_2025_104210 crossref_primary_10_1109_TGRS_2023_3320044 crossref_primary_10_1109_JIOT_2024_3510695 crossref_primary_10_1016_j_cose_2024_103740 crossref_primary_10_1007_s44443_025_00115_1 crossref_primary_10_1109_TSP_2023_3340028 crossref_primary_10_3390_electronics12224633 crossref_primary_10_1109_COMST_2024_3434510 crossref_primary_10_1145_3760788 crossref_primary_10_1145_3651307 crossref_primary_10_1109_JIOT_2023_3325634 crossref_primary_10_1007_s42979_025_03934_1 crossref_primary_10_1109_TIFS_2025_3539964 crossref_primary_10_1109_TNNLS_2022_3213187 crossref_primary_10_1016_j_automatica_2023_111264 crossref_primary_10_1109_TETC_2024_3474484 crossref_primary_10_1109_ACCESS_2025_3602217 crossref_primary_10_1016_j_future_2024_04_048 crossref_primary_10_1109_JSEN_2023_3287580 crossref_primary_10_1109_TIFS_2025_3579290 crossref_primary_10_1016_j_ifacol_2025_07_065 crossref_primary_10_1016_j_comnet_2025_111187 crossref_primary_10_1016_j_comnet_2025_111626 crossref_primary_10_32604_cmes_2024_048932 crossref_primary_10_3390_math12060901 crossref_primary_10_1109_ACCESS_2024_3431031 crossref_primary_10_1109_TNSM_2024_3514212 crossref_primary_10_32604_cmc_2024_058656 crossref_primary_10_1016_j_jpdc_2024_104918 crossref_primary_10_1016_j_ins_2023_119743 crossref_primary_10_1109_ACCESS_2025_3586416 crossref_primary_10_1016_j_jobe_2025_113223 crossref_primary_10_1016_j_comnet_2024_110383 crossref_primary_10_1109_ACCESS_2023_3284976 crossref_primary_10_1145_3706112 crossref_primary_10_1109_JIOT_2025_3528013 crossref_primary_10_1109_TFUZZ_2024_3412983 crossref_primary_10_1109_TIFS_2024_3378006 crossref_primary_10_1016_j_knosys_2024_112029 crossref_primary_10_1016_j_knosys_2024_111851 crossref_primary_10_3390_math11010214 crossref_primary_10_1007_s40747_023_00978_9 crossref_primary_10_1109_JIOT_2024_3421602 crossref_primary_10_1002_mp_16964 crossref_primary_10_1016_j_comcom_2025_108272 crossref_primary_10_1109_JIOT_2024_3452717 crossref_primary_10_1109_TNSM_2023_3336067 crossref_primary_10_1016_j_neucom_2023_127225 crossref_primary_10_1109_TSUSC_2024_3379440 crossref_primary_10_1016_j_ins_2024_120475 crossref_primary_10_1109_TIFS_2025_3576008 crossref_primary_10_3390_electronics13081540 crossref_primary_10_1109_ACCESS_2025_3567487 crossref_primary_10_1177_14727978251366544 crossref_primary_10_1007_s00521_025_11475_0 crossref_primary_10_1007_s10586_024_04894_6 crossref_primary_10_1109_TC_2023_3257510 crossref_primary_10_1109_TCBB_2022_3185395 crossref_primary_10_1002_int_22993 crossref_primary_10_1007_s40747_024_01664_0 crossref_primary_10_1109_TIFS_2025_3536777 crossref_primary_10_1016_j_patcog_2024_110976 crossref_primary_10_1016_j_compbiomed_2024_108905 crossref_primary_10_1109_ACCESS_2025_3605870 crossref_primary_10_1109_TVT_2024_3524834 crossref_primary_10_1109_ACCESS_2024_3413069 crossref_primary_10_1109_ACCESS_2024_3493946 crossref_primary_10_1007_s11227_025_07619_7 crossref_primary_10_1016_j_ins_2022_10_113 crossref_primary_10_1016_j_neucom_2025_131580 crossref_primary_10_1109_TNNLS_2025_3563537 crossref_primary_10_1007_s13042_025_02750_6 crossref_primary_10_1016_j_future_2024_04_009 crossref_primary_10_1016_j_compbiomed_2025_111060 crossref_primary_10_1007_s42979_024_03462_4 crossref_primary_10_4018_JCIT_333858 crossref_primary_10_1016_j_jnca_2023_103714 crossref_primary_10_1109_TPAMI_2025_3581555 crossref_primary_10_1109_TC_2023_3293731 crossref_primary_10_1109_JSAC_2024_3431526 crossref_primary_10_1109_JIOT_2024_3523293 crossref_primary_10_3390_fi16100374 crossref_primary_10_3390_electronics12051190 crossref_primary_10_1016_j_eswa_2025_127636 crossref_primary_10_3390_s23177358 crossref_primary_10_1016_j_ins_2023_119784 crossref_primary_10_3390_jcp3020010 crossref_primary_10_1109_TDSC_2025_3550330 crossref_primary_10_1016_j_inffus_2024_102251 crossref_primary_10_1109_TMC_2024_3407792 crossref_primary_10_1007_s11227_025_07449_7 crossref_primary_10_1109_TIFS_2025_3587185 crossref_primary_10_1016_j_engappai_2024_109464 crossref_primary_10_1109_TIFS_2023_3345171 crossref_primary_10_1016_j_engappai_2025_110724 crossref_primary_10_1109_ACCESS_2025_3580198 crossref_primary_10_1109_TSMC_2025_3551264 crossref_primary_10_1016_j_future_2024_05_009 crossref_primary_10_1016_j_jnca_2025_104280 crossref_primary_10_1109_TMC_2024_3427709 crossref_primary_10_1109_JIOT_2024_3462674 crossref_primary_10_1109_ACCESS_2025_3593953 crossref_primary_10_1007_s11704_024_3767_z crossref_primary_10_1177_14727978241299490 crossref_primary_10_1002_cpe_8084 crossref_primary_10_1016_j_comnet_2025_111442 crossref_primary_10_1007_s42952_023_00249_w crossref_primary_10_1016_j_sysarc_2024_103258 crossref_primary_10_1016_j_adhoc_2025_103934 crossref_primary_10_3390_s22082991 crossref_primary_10_1016_j_procs_2023_09_021 crossref_primary_10_1109_TIFS_2024_3420135 crossref_primary_10_1109_TNET_2023_3324697 crossref_primary_10_1016_j_comnet_2024_110730 crossref_primary_10_1109_TNNLS_2022_3216981 crossref_primary_10_3390_electronics12040842 crossref_primary_10_1016_j_knosys_2024_111660 crossref_primary_10_1109_TIFS_2025_3577498 crossref_primary_10_1016_j_aej_2025_05_008 crossref_primary_10_1109_TII_2024_3435449 crossref_primary_10_1145_3678182 crossref_primary_10_1145_3725814 crossref_primary_10_3390_electronics12102287 crossref_primary_10_3390_math12182932 crossref_primary_10_1016_j_inffus_2025_103569 crossref_primary_10_3390_app13053135 crossref_primary_10_1016_j_ins_2023_119808 crossref_primary_10_1109_TIT_2022_3192874 crossref_primary_10_1109_TBDATA_2023_3237397 crossref_primary_10_1109_JIOT_2023_3348117 crossref_primary_10_1016_j_inffus_2022_09_011 crossref_primary_10_1109_ACCESS_2024_3363884 crossref_primary_10_1016_j_asoc_2024_111764 crossref_primary_10_1016_j_future_2024_01_007 crossref_primary_10_1002_cpe_70219 crossref_primary_10_1016_j_ins_2024_120760 crossref_primary_10_1109_ACCESS_2022_3226245 crossref_primary_10_1109_TCE_2024_3386932 crossref_primary_10_1145_3669902 crossref_primary_10_1016_j_jnca_2025_104181 crossref_primary_10_1109_TDSC_2022_3204535 crossref_primary_10_1007_s10115_025_02339_z crossref_primary_10_1109_TSIPN_2024_3379844 crossref_primary_10_3390_e27060553 crossref_primary_10_1016_j_phycom_2025_102634 crossref_primary_10_1186_s40537_025_01165_y crossref_primary_10_1109_OJCOMS_2025_3558672 crossref_primary_10_1109_IOTM_001_2100192 crossref_primary_10_1049_ipr2_70201 crossref_primary_10_3390_electronics12194179 crossref_primary_10_1109_TSP_2025_3547787 crossref_primary_10_1109_TPAMI_2024_3418862 crossref_primary_10_1109_JIOT_2025_3584818 crossref_primary_10_1016_j_jmsy_2024_09_005 crossref_primary_10_1098_rsta_2024_0235 crossref_primary_10_1109_TWC_2023_3326438 crossref_primary_10_1016_j_energy_2023_129847 crossref_primary_10_1063_5_0293257 crossref_primary_10_1016_j_eswa_2025_128408 crossref_primary_10_1109_TCNS_2025_3527255 crossref_primary_10_1371_journal_pone_0327733 crossref_primary_10_1109_TPAMI_2022_3162397 crossref_primary_10_3390_electronics13183716 crossref_primary_10_3390_technologies13090409 crossref_primary_10_1145_3722562 crossref_primary_10_1109_OJSP_2025_3578273 crossref_primary_10_1016_j_iot_2023_100793 crossref_primary_10_1016_j_comnet_2025_111234 crossref_primary_10_1016_j_ins_2024_121274 crossref_primary_10_1016_j_engappai_2023_107166 crossref_primary_10_1016_j_eswa_2023_122861 crossref_primary_10_3390_ai6090218 crossref_primary_10_1016_j_comnet_2025_111479 crossref_primary_10_1109_LSP_2024_3434467 crossref_primary_10_1109_TIFS_2024_3402113 crossref_primary_10_1109_TMC_2025_3571058 crossref_primary_10_1016_j_future_2025_107831 crossref_primary_10_1109_ACCESS_2023_3269980 crossref_primary_10_1109_JIOT_2023_3325443 crossref_primary_10_1016_j_enbuild_2023_113449 crossref_primary_10_3390_electronics13111999 crossref_primary_10_1016_j_future_2023_03_042 crossref_primary_10_1016_j_jpdc_2024_104994 crossref_primary_10_1016_j_neucom_2023_126897 crossref_primary_10_1186_s42400_024_00341_6 crossref_primary_10_3390_drones9080528 crossref_primary_10_3390_electronics14081651 crossref_primary_10_1016_j_inffus_2025_103248 crossref_primary_10_1109_TDSC_2024_3372634 crossref_primary_10_1109_TKDE_2024_3420778 crossref_primary_10_1109_JETCAS_2024_3477348 crossref_primary_10_1109_TCE_2024_3471921 crossref_primary_10_1109_ACCESS_2024_3508030 crossref_primary_10_1016_j_knosys_2024_112484 crossref_primary_10_1109_TSP_2024_3514802 crossref_primary_10_3390_electronics12183938 crossref_primary_10_1109_TIFS_2025_3533899 crossref_primary_10_1109_ACCESS_2023_3269880 crossref_primary_10_1109_TII_2023_3329688 crossref_primary_10_1016_j_jisa_2025_104144 crossref_primary_10_1016_j_comnet_2023_109712 crossref_primary_10_1109_TCSII_2023_3249461 crossref_primary_10_1109_TIFS_2024_3451359 crossref_primary_10_1109_TNSE_2024_3397719 crossref_primary_10_1109_TSP_2023_3277211 crossref_primary_10_1109_TSC_2023_3332703 crossref_primary_10_1049_cth2_12761 crossref_primary_10_1109_TNET_2023_3301972 crossref_primary_10_3390_sym17091398 crossref_primary_10_1016_j_knosys_2025_114253 crossref_primary_10_1109_JIOT_2024_3489032 crossref_primary_10_1145_3735125 crossref_primary_10_1016_j_knosys_2025_114494 crossref_primary_10_1007_s10791_025_09676_1 crossref_primary_10_1016_j_iotcps_2023_04_001 crossref_primary_10_1109_TIFS_2024_3416042 crossref_primary_10_1109_TMC_2022_3178167 crossref_primary_10_1109_TDSC_2025_3542437 crossref_primary_10_1016_j_segan_2024_101537 crossref_primary_10_1007_s13042_025_02789_5 crossref_primary_10_3390_app14010351 crossref_primary_10_1109_TIFS_2025_3586491 crossref_primary_10_1007_s10462_024_10969_y crossref_primary_10_1007_s10586_025_05386_x crossref_primary_10_1109_JIOT_2024_3467275 crossref_primary_10_1109_MNET_2024_3486228 crossref_primary_10_1016_j_future_2023_01_026 crossref_primary_10_1109_TAI_2024_3524954 crossref_primary_10_1016_j_iot_2023_100875 crossref_primary_10_1016_j_procs_2025_02_064 crossref_primary_10_1109_JPROC_2023_3306773 crossref_primary_10_1016_j_jisa_2025_104127 crossref_primary_10_1109_TSP_2025_3564842 crossref_primary_10_1109_TII_2022_3165869 crossref_primary_10_1016_j_jisa_2024_103890 crossref_primary_10_1145_3727643 crossref_primary_10_3390_s23042112 crossref_primary_10_1109_JIOT_2025_3533087 crossref_primary_10_1038_s41598_025_09672_x crossref_primary_10_1007_s42514_024_00199_7 crossref_primary_10_1016_j_enbuild_2024_115164 crossref_primary_10_1016_j_neunet_2024_107016 crossref_primary_10_3390_electronics11091449 |
| Cites_doi | 10.1561/9781601988515 10.1109/CISS50987.2021.9400318 10.1109/5.726791 10.4171/JEMS/937 10.1145/3133956.3134098 10.3115/v1/D14-1162 10.1109/ACCESS.2019.2919736 10.1109/TSP.2019.2946020 10.1214/aos/1176347978 10.1609/aaai.v33i01.33011544 10.1007/978-3-030-38438-8 10.1057/palgrave.jors.2600425 10.1214/19-aos1828 10.1007/BF01584648 10.1109/MSP.2020.2975749 10.1109/TSMCC.2009.2032660 10.1109/ICDCS47774.2020.00032 10.1109/ICCV48922.2021.00487 10.1137/1.9781611975482.171 10.1561/9781680835090 10.1561/2200000083 10.1145/357172.357176 10.1214/aoms/1177703732 10.1137/100818327 10.1002/SERIES1345 10.1007/PL00011435 10.1007/978-3-319-91578-4 10.1007/11681878_14 10.1109/TSP.2017.2703664 10.3150/14-BEJ645 10.1109/TSP.2019.2893846 10.1145/3154503 10.1145/2897518.2897647 10.1007/BF01580224 10.1137/140957639 10.3150/18-bej1046 10.1016/j.jbi.2019.103291 10.1145/3372297.3417885 10.1214/17-AOS1607 10.1080/10556788.2016.1278445 10.1109/FOCS.2016.85 10.1145/1666420.1666444 10.1109/TSP.2019.2931171 10.1137/13094829X 10.1007/s10957-014-0586-7 10.1145/3133956.3133982 10.1007/978-3-642-04898-2_594 10.1109/TSP.2020.3012952 10.1162/neco.1997.9.8.1735 |
| ContentType | Journal Article |
| Copyright | Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022 |
| Copyright_xml | – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022 |
| DBID | 97E RIA RIE AAYXX CITATION 7SC 7SP 8FD JQ2 L7M L~C L~D |
| DOI | 10.1109/TSP.2022.3153135 |
| DatabaseName | IEEE Xplore (IEEE) IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Computer and Information Systems Abstracts Electronics & Communications Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional |
| DatabaseTitle | CrossRef Technology Research Database Computer and Information Systems Abstracts – Academic Electronics & Communications Abstracts ProQuest Computer Science Collection Computer and Information Systems Abstracts Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Professional |
| 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 | 1941-0476 |
| EndPage | 1154 |
| ExternalDocumentID | 10_1109_TSP_2022_3153135 9721118 |
| Genre | orig-research |
| GrantInformation_xml | – fundername: Washington Research Foundation funderid: 10.13039/100001906 – fundername: Canadian Institute for Advanced Research; CIFAR funderid: 10.13039/100007631 – fundername: NSF grantid: CCF-1740551; CCF-1703574; DMS-1839371 – fundername: JP Morgan Ph.D. Fellowship |
| GroupedDBID | -~X .DC 0R~ 29I 3EH 4.4 53G 5GY 5VS 6IK 85S 97E AAJGR AARMG AASAJ AAWTH ABAZT ABFSI ABQJQ ABVLG ACGFO ACIWK ACKIV ACNCT AENEX AETIX AGQYO AGSQL AHBIQ AI. AIBXA AJQPL AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 E.L EBS EJD F5P HZ~ H~9 ICLAB IFIPE IFJZH IPLJI JAVBF LAI MS~ O9- OCL P2P RIA RIE RNS TAE TN5 VH1 AAYXX CITATION 7SC 7SP 8FD JQ2 L7M L~C L~D |
| ID | FETCH-LOGICAL-c291t-800d178ac86f24eee3a10573ff8f27cb1a2c1584897f91eab8c8f42df9ffd7123 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 369 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000769985200008&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1053-587X |
| IngestDate | Mon Jun 30 10:16:25 EDT 2025 Sat Nov 29 04:10:55 EST 2025 Tue Nov 18 21:59:35 EST 2025 Wed Aug 27 02:10:28 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| 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-c291t-800d178ac86f24eee3a10573ff8f27cb1a2c1584897f91eab8c8f42df9ffd7123 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| ORCID | 0000-0002-1262-8466 |
| PQID | 2639931021 |
| PQPubID | 85478 |
| PageCount | 13 |
| ParticipantIDs | proquest_journals_2639931021 ieee_primary_9721118 crossref_citationtrail_10_1109_TSP_2022_3153135 crossref_primary_10_1109_TSP_2022_3153135 |
| PublicationCentury | 2000 |
| PublicationDate | 20220000 2022-00-00 20220101 |
| PublicationDateYYYYMMDD | 2022-01-01 |
| PublicationDate_xml | – year: 2022 text: 20220000 |
| PublicationDecade | 2020 |
| PublicationPlace | New York |
| PublicationPlace_xml | – name: New York |
| PublicationTitle | IEEE transactions on signal processing |
| PublicationTitleAbbrev | TSP |
| PublicationYear | 2022 |
| 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 Yang (ref3) 2018; abs/1812.02903 ref59 Nesterov (ref89) 2018 din (ref4) 2019; abs/1901.09888 ref58 Huber (ref35) 2011 ref11 ref55 ref10 ref54 Blanchard (ref49) 2017 Minsker (ref43) 2019 He (ref31) 2018; 31 Alistarh (ref53) 2018 Li (ref20) 2020 Reddi (ref26) 2021 ref50 Shakespeare (ref79) ref91 Smith (ref29) 2018; 18 ref90 ref46 Caldas (ref77); abs/1812.01097 ref45 ref48 ref47 ref42 ref86 ref41 ref85 ref44 Jain (ref72) 2017; 2 ref88 Bach (ref71) 2013 Karimireddy (ref21) 2020 Weiszfeld (ref14) 1937; 43 Bertsekas (ref28) 1997 Jain (ref73) 2017; 18 ref8 ref7 Sun (ref84) 2019; abs/1911.07963 ref9 Agarwal (ref76) 2020; abs/2003.12880 ref6 ref5 Chen (ref51) 2018 ref82 ref40 Kairouz (ref63); 139 ref80 ref34 ref37 ref75 ref30 ref74 ref33 Yin (ref52) 2018 Smith (ref18) 2017; 30 Zhang (ref83) 2017 Cohen (ref78) 2017; abs/1702.05373 ref2 Nemirovski (ref36) 1983 ref39 Avdiukhin (ref25) 2021; 139 Leblond (ref32) 2018; 19 McMahan (ref1) 2017 Wang (ref66) 2021; abs/2107.06917 Hsu (ref38) 2016; 17 Dinh (ref22) 2020; 33 ref70 ref24 ref68 ref69 ref64 Mohri (ref19) 2019; 97 Donoho (ref67) 1983; 157184 J (ref87) 2013 ref27 Bonawitz (ref13) 2019 Fallah (ref23) 2020 Balle (ref17) 2020 Gafni (ref65) 2021; abs/2103.17150 ref60 ref62 ref61 Go (ref81) 2009 |
| References_xml | – volume: abs/1812.02903 volume-title: CoRR year: 2018 ident: ref3 article-title: Applied federated learning: Improving google keyboard query suggestions – ident: ref33 doi: 10.1561/9781601988515 – ident: ref24 doi: 10.1109/CISS50987.2021.9400318 – volume: 17 start-page: 18:1 year: 2016 ident: ref38 article-title: Loss minimization and parameter estimation with heavy tails publication-title: J. Mach. Learn. Res. – ident: ref91 doi: 10.1109/5.726791 – ident: ref39 doi: 10.4171/JEMS/937 – ident: ref55 doi: 10.1145/3133956.3134098 – ident: ref82 doi: 10.3115/v1/D14-1162 – ident: ref7 doi: 10.1109/ACCESS.2019.2919736 – volume-title: Parallel and Distributed Computation: Numerical Methods year: 1997 ident: ref28 – ident: ref54 doi: 10.1109/TSP.2019.2946020 – volume: abs/2107.06917 volume-title: CoRR year: 2021 ident: ref66 article-title: A field guide to federated optimization – ident: ref70 doi: 10.1214/aos/1176347978 – volume: 139 start-page: 5201 volume-title: Proc. ICML ident: ref63 article-title: The distributed discrete Gaussian mechanism for federated learning with secure aggregation – ident: ref56 doi: 10.1609/aaai.v33i01.33011544 – volume: 33 start-page: 21394 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2020 ident: ref22 article-title: Personalized federated learning with moreau envelopes – ident: ref74 doi: 10.1007/978-3-030-38438-8 – ident: ref86 doi: 10.1057/palgrave.jors.2600425 – ident: ref40 doi: 10.1214/19-aos1828 – ident: ref57 doi: 10.1007/BF01584648 – start-page: 119 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2017 ident: ref49 article-title: Machine learning with adversaries: Byzantine tolerant gradient descent – volume: 97 start-page: 4615 volume-title: Proc. Int. Conf. Mach. Learn. year: 2019 ident: ref19 article-title: Agnostic federated learning – ident: ref27 doi: 10.1109/MSP.2020.2975749 – volume-title: Problem Complexity and Method Efficiency in Optimization year: 1983 ident: ref36 – ident: ref5 doi: 10.1109/TSMCC.2009.2032660 – start-page: 1 year: 2009 ident: ref81 article-title: Twitter sentiment classification using distant supervision – volume: 139 start-page: 425 volume-title: Proc. Int. Conf. Mach. Learn. year: 2021 ident: ref25 article-title: Federated learning under arbitrary communication patterns – ident: ref8 doi: 10.1109/ICDCS47774.2020.00032 – ident: ref9 doi: 10.1109/ICCV48922.2021.00487 – ident: ref44 doi: 10.1137/1.9781611975482.171 – ident: ref68 doi: 10.1561/9781680835090 – ident: ref2 doi: 10.1561/2200000083 – volume: 31 start-page: 4541 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2018 ident: ref31 article-title: COLA: Decentralized linear learning – ident: ref48 doi: 10.1145/357172.357176 – ident: ref34 doi: 10.1214/aoms/1177703732 – ident: ref85 doi: 10.1137/100818327 – volume: abs/1901.09888 volume-title: CoRR year: 2019 ident: ref4 article-title: Federated collaborative filtering for privacy-preserving personalized recommendation system – ident: ref10 doi: 10.1002/SERIES1345 – volume: 19 start-page: 81:1 year: 2018 ident: ref32 article-title: Improved asynchronous parallel optimization analysis for stochastic incremental methods publication-title: J. Mach. Learn. Res. – ident: ref58 doi: 10.1007/PL00011435 – volume-title: Introductory Lectures on Convex Optimization year: 2018 ident: ref89 doi: 10.1007/978-3-319-91578-4 – start-page: 2496 volume-title: Proc. 23rd Int. Conf. Artif. Intell. Statist. year: 2020 ident: ref17 article-title: Hypothesis testing interpretations and renyi differential privacy – year: 2019 ident: ref43 article-title: Uniform bounds for robust mean estimators – volume: abs/1702.05373 volume-title: CoRR year: 2017 ident: ref78 article-title: EMNIST: An extension of MNIST to handwritten letters – volume-title: Proc. Int. Conf. Learn. Representations year: 2021 ident: ref26 article-title: Adaptive federated optimization – start-page: 374 volume-title: Proc. Mach. Learn. Syst. year: 2019 ident: ref13 article-title: Towards federated learning at scale: System design – ident: ref62 doi: 10.1007/11681878_14 – volume: 2 start-page: 2:1 volume-title: Proc. Conf. Found. Softw. Technol. Theor. Comput. Sci. year: 2017 ident: ref72 article-title: A Markov chain theory approach to characterizing the minimax optimality of stochastic gradient descent (for least squares – volume: 18 start-page: 230:1 year: 2018 ident: ref29 article-title: COCOA: A general framework for communication-efficient distributed optimization publication-title: J. Mach. Learn. Res. – start-page: 902 volume-title: Proc. Int. Conf. Mach. Learn. year: 2018 ident: ref51 article-title: DRACO: Byzantine-resilient distributed training via redundant gradients – ident: ref45 doi: 10.1109/TSP.2017.2703664 – ident: ref37 doi: 10.3150/14-BEJ645 – ident: ref46 doi: 10.1109/TSP.2019.2893846 – start-page: 4618 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2018 ident: ref53 article-title: Byzantine stochastic gradient descent – ident: ref50 doi: 10.1145/3154503 – ident: ref61 doi: 10.1145/2897518.2897647 – ident: ref60 doi: 10.1007/BF01580224 – ident: ref88 doi: 10.1137/140957639 – start-page: 3557 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2020 ident: ref23 article-title: Personalized federated learning with theoretical guarantees: A model-agnostic meta-learning approach – ident: ref41 doi: 10.3150/18-bej1046 – volume: 43 start-page: 355 year: 1937 ident: ref14 article-title: On the point for which the sum of the distances to n given points is minimum publication-title: Tohoku Math. J., First Ser. – volume: abs/1812.01097 volume-title: CoRR ident: ref77 article-title: LEAF: A benchmark for federated settings – ident: ref6 doi: 10.1016/j.jbi.2019.103291 – ident: ref12 doi: 10.1145/3372297.3417885 – ident: ref69 doi: 10.1214/17-AOS1607 – volume: 30 start-page: 4424 volume-title: Proc. Adv. Neural Informat. Process. Syst. year: 2017 ident: ref18 article-title: Federated multi-task learning – ident: ref30 doi: 10.1080/10556788.2016.1278445 – ident: ref42 doi: 10.1109/FOCS.2016.85 – ident: ref64 doi: 10.1145/1666420.1666444 – start-page: 773 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2013 ident: ref71 article-title: Non-strongly-convex smooth stochastic approximation with convergence rate $O(1/n)$ – ident: ref47 doi: 10.1109/TSP.2019.2931171 – volume: abs/1911.07963 volume-title: CoRR year: 2019 ident: ref84 – start-page: 5132 volume-title: Proc. Int. Conf. Mach. Learn. year: 2020 ident: ref21 article-title: Scaffold: Stochastic controlled averaging for federated learning – ident: ref79 article-title: The complete works of William Shakespeare – ident: ref90 doi: 10.1137/13094829X – ident: ref59 doi: 10.1007/s10957-014-0586-7 – volume-title: Proc. Int. Conf. Learn. Representations year: 2017 ident: ref83 article-title: Understanding deep learning requires rethinking generalization – ident: ref11 doi: 10.1145/3133956.3133982 – volume: 157184 start-page: 57 year: 1983 ident: ref67 article-title: The notion of breakdown point publication-title: A festschrift Erich L. Lehmann – start-page: 1248 volume-title: Robust Statistics. in International Encyclopedia of Statistical Science year: 2011 ident: ref35 doi: 10.1007/978-3-642-04898-2_594 – start-page: 783 volume-title: Proc. Int. Conf. Mach. Learn. year: 2013 ident: ref87 article-title: Optimization with first-order surrogate functions – start-page: 429 volume-title: Proc. Mach. Learn. Syst. year: 2020 ident: ref20 article-title: Federated optimization in heterogeneous networks – start-page: 5636 volume-title: Proc. Int. Conf. Mach. Learn. year: 2018 ident: ref52 article-title: Byzantine-robust distributed learning: Towards optimal statistical rates – volume: 18 start-page: 223:1 year: 2017 ident: ref73 article-title: Parallelizing stochastic gradient descent for least squares regression: Mini-batching, averaging, and model misspecification publication-title: J. Mach. Learn. Res. – ident: ref75 doi: 10.1109/TSP.2020.3012952 – volume: abs/2003.12880 volume-title: CoRR year: 2020 ident: ref76 article-title: Federated residual learning – start-page: 1273 volume-title: Proc. Artif. Intell. Statist. year: 2017 ident: ref1 article-title: Communication-efficient learning of deep networks from decentralized data – volume: abs/2103.17150 volume-title: CoRR year: 2021 ident: ref65 article-title: Federated learning: A signal processing perspective – ident: ref80 doi: 10.1162/neco.1997.9.8.1735 |
| SSID | ssj0014496 |
| Score | 2.7431302 |
| Snippet | We present a novel approach to federated learning that endows its aggregation process with greater robustness to potential poisoning of local data or model... |
| SourceID | proquest crossref ieee |
| SourceType | Aggregation Database Enrichment Source Index Database Publisher |
| StartPage | 1142 |
| SubjectTerms | Agglomeration Aggregates Algorithms Collaborative work Computer vision corrupted updates Corruption data privacy distributed learning Federated learning Machine learning Natural language processing Optimization Privacy robust aggregation Robustness Servers Signal processing algorithms |
| Title | Robust Aggregation for Federated Learning |
| URI | https://ieeexplore.ieee.org/document/9721118 https://www.proquest.com/docview/2639931021 |
| Volume | 70 |
| WOSCitedRecordID | wos000769985200008&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: 1941-0476 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014496 issn: 1053-587X databaseCode: RIE dateStart: 19910101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PS8MwFH7M4UEP_pridEoPXgbWNWm6JMchDg8yhk7ZraRpMgTZZOv8-31JuyoogrcekhK-pHnfy2u-D-CqT61VnKpQm5yGzFgaKpZEYaJznhOheFkueHngo5GYTuW4Adf1XRhjjP_5zNy4R1_Lzxd67Y7Kek5pBgnxFmxx3i_vatUVA8a8FxfShThMBJ9uSpKR7E2expgIUor5Ka44b-z2FYK8p8qPjdhHl-H-_8Z1AHsViwwG5bQfQsPMj2D3m7ZgC7qPi2y9KoLBDDPqmcc_QIIaDJ16BBLMPKikVWfH8Dy8m9zeh5UvQqipJAUGlSgnXCgt-pYyHE2snFtvbK2wlOuMKKoJEgshuZXEqExoYRnNrbQ25xiqTqA5X8zNKQTSZkbFiSYxZ8wmJMPsJTLYhGncBZVuQ28DVaor0XDnXfGW-uQhkimCmzpw0wrcNnTrHu-lYMYfbVsOzLpdhWMbOpvZSKsvapVST6WcEfnZ773OYce9uzwe6UCzWK7NBWzrj-J1tbz0i-UTR0G65Q |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PS8MwFH7MKagHf01xOrUHLwPrmjRd0uMQx8Q5hk7ZraT5MQTZZOv8-03SrgqK4K2HFxreS_O-l9d8H8BlG2vNKea-UBL7RGnscxIFfiQklYhxmrcLXvp0MGDjcTyswFV5F0Yp5X4-U9f20fXy5Uws7VFZyzLNGEC8BusRITjIb2uVPQNCnBqXAQyhHzE6XjUlg7g1ehqaUhBjU6GaNeek3b6SkFNV-bEVu_zS3f3fzPZgp8CRXicP_D5U1PQAtr-xC9ag-ThLl4vM60xMTT1xEfAMRPW6lj_CQEzpFeSqk0N47t6Obnp-oYzgCxyjzKSVQCLKuGBtjYmZTcitXm-oNdOYihRxLJCBFiymOkaKp0wwTbDUsdaSmmR1BNXpbKqOwYt1qngYCRRSQnSEUlO_BMqYEGH2QS7q0Fq5KhEFbbhVr3hLXPkQxIlxbmKdmxTOrUOzHPGeU2b8YVuzziztCj_WobGKRlJ8U4sEOzBlpchPfh91AZu90UM_6d8N7k9hy74nPyxpQDWbL9UZbIiP7HUxP3cL5xMwHL4s |
| 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=Robust+Aggregation+for+Federated+Learning&rft.jtitle=IEEE+transactions+on+signal+processing&rft.au=Pillutla%2C+Krishna&rft.au=Kakade%2C+Sham+M.&rft.au=Harchaoui%2C+Zaid&rft.date=2022&rft.issn=1053-587X&rft.eissn=1941-0476&rft.volume=70&rft.spage=1142&rft.epage=1154&rft_id=info:doi/10.1109%2FTSP.2022.3153135&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_TSP_2022_3153135 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1053-587X&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1053-587X&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1053-587X&client=summon |