%PDF-1.3 1 0 obj << /Kids [ 4 0 R 5 0 R 6 0 R 7 0 R 8 0 R 9 0 R 10 0 R 11 0 R 12 0 R 13 0 R 14 0 R ] /Type /Pages /Count 11 >> endobj 2 0 obj << /Subject (Neural Information Processing Systems http\072\057\057nips\056cc\057) /Publisher (Curran Associates\054 Inc\056) /Language (en\055US) /Created (2019) /EventType (Poster) /Description-Abstract (The simplest and most widely applied method for guaranteeing differential privacy is to add instance\055independent noise to a statistic of interest that is scaled to its global sensitivity\056 However\054 global sensitivity is a worst\055case notion that is often too conservative for realized dataset instances\056 We provide methods for scaling noise in an instance\055dependent way and demonstrate that they provide greater accuracy under average\055case distributional assumptions\056 Specifically\054 we consider the basic problem of privately estimating the mean of a real distribution from i\056i\056d\056 samples\056 The standard empirical mean estimator can have arbitrarily\055high global sensitivity\056 We propose the trimmed mean estimator\054 which interpolates between the mean and the median\054 as a way of attaining much lower sensitivity on average while losing very little in terms of statistical accuracy\056 To privately estimate the trimmed mean\054 we revisit the smooth sensitivity framework of Nissim\054 Raskhodnikova\054 and Smith \050STOC 2007\051\054 which provides a framework for using instance\055dependent sensitivity\056 We propose three new additive noise distributions which provide concentrated differential privacy when scaled to smooth sensitivity\056 We provide theoretical and experimental evidence showing that our noise distributions compare favorably to others in the literature\054 in particular\054 when applied to the mean estimation problem\056) /Producer (PyPDF2) /Title (Average\055Case Averages\072 Private Algorithms for Smooth Sensitivity and Mean Estimation) /Date (2019) /ModDate (D\07220200212210316\05508\04700\047) /Published (2019) /Type (Conference Proceedings) /firstpage (181) /Book (Advances in Neural Information Processing Systems 32) /Description (Paper accepted and presented at the Neural Information Processing Systems Conference \050http\072\057\057nips\056cc\057\051) /Editors (H\056 Wallach and H\056 Larochelle and A\056 Beygelzimer and F\056 d\047Alch\351\055Buc and E\056 Fox and R\056 Garnett) /Author (Mark Bun\054 Thomas Steinke) /lastpage (191) >> endobj 3 0 obj << /Type /Catalog /Pages 1 0 R >> endobj 4 0 obj << /Contents 15 0 R /Parent 1 0 R /Resources 16 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 71 0 R ] /Type /Page >> endobj 5 0 obj << /Contents 72 0 R /Parent 1 0 R /Resources 73 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 78 0 R 79 0 R 80 0 R 81 0 R 82 0 R 83 0 R 84 0 R 85 0 R ] /Type /Page >> endobj 6 0 obj << /Contents 86 0 R /Parent 1 0 R /Resources 87 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 92 0 R 93 0 R 94 0 R 95 0 R 96 0 R 97 0 R 98 0 R ] /Type /Page >> endobj 7 0 obj << /Contents 99 0 R /Parent 1 0 R /Resources 100 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 109 0 R 110 0 R 111 0 R 112 0 R 113 0 R 114 0 R 115 0 R 116 0 R 117 0 R 118 0 R 119 0 R 120 0 R ] /Type /Page >> endobj 8 0 obj << /Contents 121 0 R /Parent 1 0 R /Resources 122 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 123 0 R 124 0 R 125 0 R 126 0 R 127 0 R 128 0 R 129 0 R 130 0 R 131 0 R 132 0 R 133 0 R 134 0 R 135 0 R 136 0 R 137 0 R 138 0 R 139 0 R 140 0 R 141 0 R 142 0 R 143 0 R 144 0 R 145 0 R 146 0 R 147 0 R 148 0 R 149 0 R 150 0 R ] /Type /Page >> endobj 9 0 obj << /Contents 151 0 R /Parent 1 0 R /Resources 152 0 R /Group 157 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 158 0 R 159 0 R 160 0 R 161 0 R 162 0 R 163 0 R 164 0 R 165 0 R 166 0 R 167 0 R 168 0 R 169 0 R 170 0 R 171 0 R 172 0 R 173 0 R 174 0 R 175 0 R ] /Type /Page >> endobj 10 0 obj << /Contents 176 0 R /Parent 1 0 R /Resources 177 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 182 0 R 183 0 R 184 0 R ] /Type /Page >> endobj 11 0 obj << /Contents 185 0 R /Parent 1 0 R /Resources 186 0 R /Group 157 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 191 0 R 192 0 R 193 0 R 194 0 R 195 0 R 196 0 R 197 0 R ] /Type /Page >> endobj 12 0 obj << /Contents 198 0 R /Parent 1 0 R /Resources 199 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 200 0 R 201 0 R 202 0 R 203 0 R 204 0 R 205 0 R 206 0 R 207 0 R 208 0 R 209 0 R ] /Type /Page >> endobj 13 0 obj << /Contents 210 0 R /Parent 1 0 R /Resources 211 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 212 0 R 213 0 R 214 0 R 215 0 R ] /Type /Page >> endobj 14 0 obj << /Contents 216 0 R /Parent 1 0 R /Type /Page /Resources 217 0 R /MediaBox [ 0 0 612 792 ] >> endobj 15 0 obj << /Length 3591 /Filter /FlateDecode >> stream xڭZsBFOO,|$әirm:%s"|_vm'g, o_z7l,θ+3(d#<ӛ]_4\݆Ko4s-7͏߮9c]q_n}y]?J]W1}UñݵQ]|F[ǹиMݶޔM_ HqhuY4y?Tu1Tmc^}~{~&Dq"L~m sN洹 D\%"ʔ&a2I~]tpף`іไO؊u-2 yidoNhp4VRiԻl7amg&*VI Ƕ.nj[e: ߾~_ez~"q\YaTkFW`OTM9q!O<~r f+g ͓8Uʩϧ̷i =T|K`߁:f(@8r@P-ؕP'|yaF7Z``3g2E+s|8pjhj幄̀S¯b~89RC١d̚cA8YYR5Co7_r&!vge :( cI (G_F4%YIVrX,.C ,GFۮ{2kZv-Pm pa&CUaũJPѡ_[8~Θ[ZC%HS7C:r|*cIpSM$Y`Ij ~{mߏ]hDĂ#F/sj;(yWFc(\=",>epb9' V`Q47]# 1.<@ASY:ky`4|* rU1| Nw][SW!&),$hnbIPt\u_{OM@=ޱ~,svRWu`!nKP"Aӑi0(nmO`p~UץĽWAG F<'GvPZ<2LΗX' YƵ*h^'~Lf _2(p,PT Y?Zn=٥4CPXlUnhcu*qީNeBL8q5<⅞a8ѕW E-:^ʩ%6ho V~M˙%Qg_yuC#` pA.̀p%E
pDz\#l fƜ%VqnەlLࡊ]DO1u%WH.G#!S棅,jG$wa_hΝ? bg-oL
" $bF]sD#@0v1
5
%AW4Q͝3'mR"1@ײ0l2H߂nƋL)4+ #&$Dx*w_aᝍb>>_CJ<#-P<) $K&TfY7PcrD9\OUNK$8\WF4x]'*_10YAtw7"m`f<\VnlmX=W
įpAT$>YEk%݄3Nr{3*Y;?,x0/Tn_P\R mҩo} o40b>m`}(Sv K^mzc6x$@Uܜk$\4
Sss ;C7*,Z%\g[طk5ӫ%Dʘ#@9Hb|(QQƾ<$yxBy9\*kFޭ*Շb"ũW<څOkG6ȁTNbLi2``}*PQv՜>\43#gHe<}N,yih?FW蟃A3?~
)IB/)o|sFXD)0 OyYs0`IQzM w3*E攚2;=]2z!SfMX^EaV;88|xNm
Xx.p_rH&+(5SE9jF_ԛOE9M:HԻrk+2SҶη8ҽjn$̟c܍w/;Iݜ5TM,A88
erx}qlw8<+O9O8jTb9Oi,Qޔ0,͏{a'w&O1 +x0 g$)--!RLΗ; ?̄qlX