[BCÖ20] Provides lower bounds for parameter estimation under $\ell_2$ loss for interactive protocols (blackboard model) under local privacy, and instantiate it to obtain tight bounds for Gaussian mean estimation, (sparse) Bernoulli mean estimation, and discrete distribution estimation. As in [BHÖ19], the lower bound framework is based on a Cramér–Rao/van Trees-type approach.
Paper bibtex 2 downloads @article{BCO20,
author = {Leighton P. Barnes and
Wei{-}Ning Chen and
Ayfer {\"{O}}zg{\"{u}}r},
title = {Fisher information under local differential privacy},
journal = {CoRR},
volume = {abs/2005.10783},
url = {https://arxiv.org/abs/2005.10783},
year = {2020},
bibbase_note = {<div class="well well-small bibbase"><span class="bluecite">[BCÖ20]</span> Provides lower bounds for parameter estimation under $\ell_2$ loss for interactive protocols (blackboard model) under local privacy, and instantiate it to obtain tight bounds for Gaussian mean estimation, (sparse) Bernoulli mean estimation, and discrete distribution estimation. As in [BHÖ19], the lower bound framework is based on a Cramér–Rao/van Trees-type approach.},
note = {Accepted to the IEEE Journal on Selected Areas in Information Theory (JSAIT).</div>}
}
Downloads: 2
{"_id":"twoAN9zx434kn7xgw","bibbaseid":"barnes-chen-zgr-fisherinformationunderlocaldifferentialprivacy-2020","authorIDs":[],"author_short":["Barnes, L. P.","Chen, W.","Özgür, A."],"bibdata":{"bibtype":"article","type":"article","author":[{"firstnames":["Leighton","P."],"propositions":[],"lastnames":["Barnes"],"suffixes":[]},{"firstnames":["Wei-Ning"],"propositions":[],"lastnames":["Chen"],"suffixes":[]},{"firstnames":["Ayfer"],"propositions":[],"lastnames":["Özgür"],"suffixes":[]}],"title":"Fisher information under local differential privacy","journal":"CoRR","volume":"abs/2005.10783","url":"https://arxiv.org/abs/2005.10783","year":"2020","bibbase_note":"<div class=\"well well-small bibbase\"><span class=\"bluecite\">[BCÖ20]</span> Provides lower bounds for parameter estimation under $\\ell_2$ loss for interactive protocols (blackboard model) under local privacy, and instantiate it to obtain tight bounds for Gaussian mean estimation, (sparse) Bernoulli mean estimation, and discrete distribution estimation. As in [BHÖ19], the lower bound framework is based on a Cramér–Rao/van Trees-type approach.","note":"Accepted to the IEEE Journal on Selected Areas in Information Theory (JSAIT).</div>","bibtex":"@article{BCO20,\n author = {Leighton P. Barnes and\n Wei{-}Ning Chen and\n Ayfer {\\\"{O}}zg{\\\"{u}}r},\n title = {Fisher information under local differential privacy},\n journal = {CoRR},\n volume = {abs/2005.10783},\n url = {https://arxiv.org/abs/2005.10783},\n year = {2020},\n bibbase_note = {<div class=\"well well-small bibbase\"><span class=\"bluecite\">[BCÖ20]</span> Provides lower bounds for parameter estimation under $\\ell_2$ loss for interactive protocols (blackboard model) under local privacy, and instantiate it to obtain tight bounds for Gaussian mean estimation, (sparse) Bernoulli mean estimation, and discrete distribution estimation. As in [BHÖ19], the lower bound framework is based on a Cramér–Rao/van Trees-type approach.},\n note = {Accepted to the IEEE Journal on Selected Areas in Information Theory (JSAIT).</div>}\n}\n\n","author_short":["Barnes, L. P.","Chen, W.","Özgür, A."],"key":"BCO20","id":"BCO20","bibbaseid":"barnes-chen-zgr-fisherinformationunderlocaldifferentialprivacy-2020","role":"author","urls":{"Paper":"https://arxiv.org/abs/2005.10783"},"metadata":{"authorlinks":{}},"downloads":2,"html":""},"bibtype":"article","biburl":"http://www.cs.columbia.edu/~ccanonne/tutorial-focs2020/publications-annotated.bib","creationDate":"2020-11-08T21:08:20.198Z","downloads":2,"keywords":[],"search_terms":["fisher","information","under","local","differential","privacy","barnes","chen","özgür"],"title":"Fisher information under local differential privacy","year":2020,"dataSources":["EYbwRcMzLdr2qsboR","SK4pKeSrqk4BzgaZs"]}