Optimal Accuracy-Privacy Trade-Off for Secure Computations
- Submitting institution
-
Imperial College of Science, Technology and Medicine
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2311
- Type
- D - Journal article
- DOI
-
10.1109/TIT.2018.2886458
- Title of journal
- IEEE Transactions on Information Theory
- Article number
- -
- First page
- 3165
- Volume
- 65
- Issue
- 5
- ISSN
- 0018-9448
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
10.1109/TIT.2018.2886458
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
-
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Applications, e.g. in healthcare, compute over private inputs where outputs inescapably leak information. This work provides the first true understanding of such leakage and develops a mechanism that optimally trades off privacy with hard guarantees on output precision, where standard differential privacy does not work well. Our PETS'20 paper (https://doi.org/10.2478/popets-2020-0053) demonstrated alternative practical algorithms based on this mechanism, and won the Andreas Pfitzmann Best Student Paper award (https://petsymposium.org/student-paper-award.php). The citation recognised that this work will "spark a lot of thoughts and discussions in the area of privacy”.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
