Mean Estimation with User-level Privacy under Data Heterogeneity
A key challenge in many modern data analysis tasks is that user data are heterogeneous. Different users may possess vastly different numbers of data points. More importantly, it cannot be assumed that all users sample from the same underlying distribution. This is true, for example in language data,...
Saved in:
| Main Authors | , , , |
|---|---|
| Format | Journal Article |
| Language | English |
| Published |
28.07.2023
|
| Subjects | |
| Online Access | Get full text |
Cover
Loading…
| Abstract | A key challenge in many modern data analysis tasks is that user data are
heterogeneous. Different users may possess vastly different numbers of data
points. More importantly, it cannot be assumed that all users sample from the
same underlying distribution. This is true, for example in language data, where
different speech styles result in data heterogeneity. In this work we propose a
simple model of heterogeneous user data that allows user data to differ in both
distribution and quantity of data, and provide a method for estimating the
population-level mean while preserving user-level differential privacy. We
demonstrate asymptotic optimality of our estimator and also prove general lower
bounds on the error achievable in the setting we introduce. |
|---|---|
| AbstractList | A key challenge in many modern data analysis tasks is that user data are
heterogeneous. Different users may possess vastly different numbers of data
points. More importantly, it cannot be assumed that all users sample from the
same underlying distribution. This is true, for example in language data, where
different speech styles result in data heterogeneity. In this work we propose a
simple model of heterogeneous user data that allows user data to differ in both
distribution and quantity of data, and provide a method for estimating the
population-level mean while preserving user-level differential privacy. We
demonstrate asymptotic optimality of our estimator and also prove general lower
bounds on the error achievable in the setting we introduce. |
| Author | Feldman, Vitaly McMillan, Audra Cummings, Rachel Talwar, Kunal |
| Author_xml | – sequence: 1 givenname: Rachel surname: Cummings fullname: Cummings, Rachel – sequence: 2 givenname: Vitaly surname: Feldman fullname: Feldman, Vitaly – sequence: 3 givenname: Audra surname: McMillan fullname: McMillan, Audra – sequence: 4 givenname: Kunal surname: Talwar fullname: Talwar, Kunal |
| BackLink | https://doi.org/10.48550/arXiv.2307.15835$$DView paper in arXiv |
| BookMark | eNotj7FOwzAURT3AQAsfwIR_ICF-juNmKyqFIhXBUObo2XktllIHOSaQv8cUhqO7XF3dM2NnvvfE2LUo8nKhVHGL4duNOchC50ItpLpgy2dCz9dDdEeMrvf8y8V3_jZQyDoaqeOvwY1oJ_7pWwr8HiPyDUUK_YE8uThdsvM9dgNd_eec7R7Wu9Um2748Pq3uthlWWmVg5V4orEFUwrQIFiTI2litSyNNBYBtapTaitLYNmFqTQmUIlUNyDm7-Zs9KTQfIf0NU_Or0pxU5A-0rUXn |
| ContentType | Journal Article |
| Copyright | http://creativecommons.org/licenses/by/4.0 |
| Copyright_xml | – notice: http://creativecommons.org/licenses/by/4.0 |
| DBID | AKY EPD GOX |
| DOI | 10.48550/arxiv.2307.15835 |
| DatabaseName | arXiv Computer Science arXiv Statistics arXiv.org |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: GOX name: arXiv.org url: http://arxiv.org/find sourceTypes: Open Access Repository |
| DeliveryMethod | fulltext_linktorsrc |
| ExternalDocumentID | 2307_15835 |
| GroupedDBID | AKY EPD GOX |
| ID | FETCH-LOGICAL-a675-2c3f15a92161bda2c23239bc774b3b622adc3f47c14bcd4bcb97eb97a31c23b23 |
| IEDL.DBID | GOX |
| IngestDate | Mon Jan 08 05:48:55 EST 2024 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | false |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-a675-2c3f15a92161bda2c23239bc774b3b622adc3f47c14bcd4bcb97eb97a31c23b23 |
| OpenAccessLink | https://arxiv.org/abs/2307.15835 |
| ParticipantIDs | arxiv_primary_2307_15835 |
| PublicationCentury | 2000 |
| PublicationDate | 2023-07-28 |
| PublicationDateYYYYMMDD | 2023-07-28 |
| PublicationDate_xml | – month: 07 year: 2023 text: 2023-07-28 day: 28 |
| PublicationDecade | 2020 |
| PublicationYear | 2023 |
| Score | 1.8874836 |
| SecondaryResourceType | preprint |
| Snippet | A key challenge in many modern data analysis tasks is that user data are
heterogeneous. Different users may possess vastly different numbers of data
points.... |
| SourceID | arxiv |
| SourceType | Open Access Repository |
| SubjectTerms | Computer Science - Cryptography and Security Computer Science - Data Structures and Algorithms Computer Science - Learning Statistics - Machine Learning |
| Title | Mean Estimation with User-level Privacy under Data Heterogeneity |
| URI | https://arxiv.org/abs/2307.15835 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwdZ07T8MwEIBPbScWBAJUnvLAakhs5-ENBC0RUoGhlbJFPtuVKqGCQlrBv-ecBMHC4Az2LbnTxV_sewBcShSxtbnlOnKSK5emHJH8SqpMLYV1uWlbJ8ye0mKhHsukHAD7yYUx9edq29UHxo_rEKV8FSdECUMYChFCth6ey-5ysi3F1cv_yhFjtlN_NonpHuz2dMduO3Psw8CvD-Bm5s2aTciXujRBFs4-2YJsz19DyA57qVdbY79YyOeq2b1pDCtClMobGdcTJR_CfDqZ3xW8b1zADfE3F1Yu48RoQTSFzghL1CI1WiItlJgKYRxJqMzGCq2jgTrzNIyMSRSFPIIR_fv7MTBl6fuTORNpK5VTGkPCY-SJaVDl9DyGcfu61XtXm6IKmqhaTZz8v3QKO6FrejiiFPkZjJp6489pb23wolXwN739eac |
| link.rule.ids | 228,230,783,888 |
| linkProvider | Cornell University |
| 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=Mean+Estimation+with+User-level+Privacy+under+Data+Heterogeneity&rft.au=Cummings%2C+Rachel&rft.au=Feldman%2C+Vitaly&rft.au=McMillan%2C+Audra&rft.au=Talwar%2C+Kunal&rft.date=2023-07-28&rft_id=info:doi/10.48550%2Farxiv.2307.15835&rft.externalDocID=2307_15835 |