Abstract | ||
---|---|---|
We present and evaluate
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Deep Private-Feature Extractor (DPFE)</italic>
, a deep model which is trained and evaluated based on information theoretic constraints. Using the selective exchange of information between a user's device and a service provider,
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPFE</italic>
enables the user to prevent certain sensitive information from being shared with a service provider, while allowing them to extract approved information using their model. We introduce and utilize the
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">log-rank</italic>
privacy, a novel measure to assess the effectiveness of
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPFE</italic>
in removing sensitive information and compare different models based on their accuracy-privacy trade-off. We then implement and evaluate the performance of
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPFE</italic>
on smartphones to understand its complexity, resource demands, and efficiency trade-offs. Our results on benchmark image datasets demonstrate that under moderate resource utilization,
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPFE</italic>
can achieve high accuracy for primary tasks while preserving the privacy of sensitive information. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/TKDE.2018.2878698 | IEEE Transactions on Knowledge and Data Engineering |
Keywords | DocType | Volume |
Data privacy,Feature extraction,Privacy,Data models,Task analysis,Training | Journal | 32 |
Issue | ISSN | Citations |
1 | 1041-4347 | 4 |
PageRank | References | Authors |
0.41 | 0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Seyed Ali Ossia | 1 | 23 | 2.09 |
Ali Taheri | 2 | 30 | 2.18 |
Ali Shahin Shamsabadi | 3 | 21 | 5.12 |
Kleomenis Katevas | 4 | 39 | 5.89 |
Hamed Haddadi | 5 | 10 | 2.91 |
Hamid Reza Rabiee | 6 | 79 | 11.48 |