slides/slides07.tex
changeset 303 0397d557c362
parent 302 17890d4b0688
child 304 9d1cf03d1b32
equal deleted inserted replaced
302:17890d4b0688 303:0397d557c362
   337 \begin{frame}[c]
   337 \begin{frame}[c]
   338 \frametitle{Re-identification Attacks}
   338 \frametitle{Re-identification Attacks}
   339 
   339 
   340 
   340 
   341 \begin{itemize}
   341 \begin{itemize}
   342 \item in 1990 insurance databases were made public with names removed, but  birth dates, 
   342 \item in 1990 medical databases were made public with names removed, but  birth dates, 
   343 gender, ZIP-code were retained\medskip
   343 gender, ZIP-code were retained\medskip
   344 \item could be cross referenced with public voter registration data in order to find out what the
   344 \item could be cross referenced with public voter registration data in order to find out what the
   345 medical record of the governor of Massachusetts was
   345 medical record of the governor of Massachusetts was (in 1997 Latanya Sweeney)
   346 \end{itemize}
   346 \end{itemize}
   347 
   347 
   348 \end{frame}}
   348 \end{frame}}
   349 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   349 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   350 
   350 
   368 \item disclosure information needs to be retained for 5 years
   368 \item disclosure information needs to be retained for 5 years
   369 \end{itemize}
   369 \end{itemize}
   370 
   370 
   371 \end{frame}}
   371 \end{frame}}
   372 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   372 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   373 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   373 
   374 \mode<presentation>{
   374 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
       
   375 \begin{frame}[c]
       
   376 \frametitle{AOL Search Queries}
       
   377 
       
   378 \begin{itemize}
       
   379 \item In 2006, AOL published 20 million Web search queries 
       
   380   collected of 650,000 users (names had been deleted)\medskip
       
   381   
       
   382 \item within days an old lady, Thelma Arnold, from Lilburn
       
   383   Georgia was identified as user No. 4417749
       
   384   
       
   385 \item some of the queries that gave her away:  
       
   386 \begin{itemize}
       
   387 \item landscapers in Lilburn, Ga
       
   388 \item 60 single men
       
   389 \item nicotine effects on the body
       
   390 \item \ldots 
       
   391 \end{itemize}  
       
   392 \end{itemize}
       
   393 
       
   394 \end{frame}
       
   395 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
       
   396 
       
   397 
       
   398 
       
   399 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   375 \begin{frame}<2>[c]
   400 \begin{frame}<2>[c]
   376 \frametitle{\large How to Safely Disclose Information?}
   401 \frametitle{\large How to Safely Disclose Information?}
   377 
   402 
   378 \only<1>{
   403 \only<1>{
   379 \begin{itemize}
   404 \begin{itemize}
   391 (GWAS was a public database of gene-frequency studies linked to diseases;
   416 (GWAS was a public database of gene-frequency studies linked to diseases;
   392 you only needed partial DNA information  in order
   417 you only needed partial DNA information  in order
   393 to identify whether an individual was part of the study --- DB closed in 2008) 
   418 to identify whether an individual was part of the study --- DB closed in 2008) 
   394 \end{itemize}}
   419 \end{itemize}}
   395 
   420 
   396 \end{frame}}
   421 \end{frame}
   397 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   422 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
   398 
   423 
   399 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   424 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   400 \mode<presentation>{
   425 \mode<presentation>{
   401 \begin{frame}<2>[c]
   426 \begin{frame}<2>[c]
   453 \item \bl{$f(x)$} can be released, if \bl{$f$} is insensitive to
   478 \item \bl{$f(x)$} can be released, if \bl{$f$} is insensitive to
   454 individual entries  \bl{$x_1, \ldots, x_n$}\\
   479 individual entries  \bl{$x_1, \ldots, x_n$}\\
   455 \item Intuition: whatever is learned from the dataset would be learned regardless of whether
   480 \item Intuition: whatever is learned from the dataset would be learned regardless of whether
   456 \bl{$x_i$} participates\bigskip\pause 
   481 \bl{$x_i$} participates\bigskip\pause 
   457 
   482 
   458 \item Noised needed in order to prevent queries:\\ Christian's salary $=$ 
   483 \item Noise needed in order to prevent queries:\\ Christian's salary $=$ 
   459 \begin{center}
   484 \begin{center}
   460 \bl{\large$\Sigma$} all staff $-$  \bl{\large$\Sigma$} all staff $\backslash$ Christian
   485 \bl{\large$\Sigma$} all staff $-$  \bl{\large$\Sigma$} all staff $\backslash$ Christian
   461 \end{center} 
   486 \end{center} 
   462 \end{itemize}
   487 \end{itemize}
   463 
   488