東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

A first principle approach toward da...

Lin, Bing-Rong.

FindBook

Google Book

Amazon

博客來

A first principle approach toward data privacy and utility.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	A first principle approach toward data privacy and utility./
作者:	Lin, Bing-Rong.
面頁冊數:	130 p.
附註:	Source: Dissertation Abstracts International, Volume: 75-11(E), Section: B.
Contained By:	Dissertation Abstracts International75-11B(E).
標題:	Computer Science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3583379
ISBN:	9781321148091

A first principle approach toward data privacy and utility.
Lin, Bing-Rong.

A first principle approach toward data privacy and utility. - 130 p.

Source: Dissertation Abstracts International, Volume: 75-11(E), Section: B.

Thesis (Ph.D.)--The Pennsylvania State University, 2014.

Individual data are collected by government and online service providers such as social networks, search engines and shopping websites. The huge amount of data about users' activities and personal information are collected and analyzed to improve the quality of service or to serve as scientific research data. However, the individual's privacy may not be protected.

ISBN: 9781321148091Subjects--Topical Terms:

626642
Computer Science.

A first principle approach toward data privacy and utility.
LDR:04000nmm a2200313 4500 001 2056210
005 20150505071904.5
008 170521s2014 ||||||||||||||||| ||eng d
020 $a 9781321148091
035 $a (MiAaPQ)AAI3583379
035 $a AAI3583379
040 $a MiAaPQ $c MiAaPQ
100 1 $a Lin, Bing-Rong. $3 3169958
245 1 2 $a A first principle approach toward data privacy and utility.
300 $a 130 p.
500 $a Source: Dissertation Abstracts International, Volume: 75-11(E), Section: B.
500 $a Adviser: Daniel Kifer.
502 $a Thesis (Ph.D.)--The Pennsylvania State University, 2014.
520 $a Individual data are collected by government and online service providers such as social networks, search engines and shopping websites. The huge amount of data about users' activities and personal information are collected and analyzed to improve the quality of service or to serve as scientific research data. However, the individual's privacy may not be protected.
520 $a The protection of privacy relies on the privacy definitions. Privacy definitions act as a contract defines the behavior of algorithms on processing sensitive data and producing non-sensitive sanitized data. Most of privacy definitions are constructed from intuition and intuition alone leads us astray. Examples include the release of AOL search log and Net ix competition dataset. The released data were anonymized based on intuition but, only a few days later, the journalist and researchers figure out means to identify people in the dataset. One goal of this thesis is to provide a systematic approach to extract semantic guarantees of privacy definitions.
520 $a Most of privacy definitions are constructed from intuition. In most cases, it is not clear what these privacy definitions actually guarantee. To the best of our knowledge, we present the first general framework for extracting semantic guarantees from privacy definitions. The privacy guarantees we can extract are Bayesian in nature and deal with changes in an attacker's beliefs. The framework can be applied to privacy definitions or even to individual algorithms to identify the types of inferences they defend against. We illustrate the use of our framework with analyses of several definitions and algorithms for which we can derive previously unknown semantics.
520 $a The other goal of this thesis is to systematically study utility measures via the first principle approach. The privacy definitions constrain the behavior of privacy algorithms. The utility measures are designed to help to choose which privacy algorithm to be used to process the sensitive data. In statistical privacy, utility refers to two concepts: information preservation { how much statistical information is retained by a sanitizing algorithm, and usability { how (and with how much difficulty) does one extract this information to build statistical models, answer queries, etc. We analyze the information-preserving properties of utility measures with utility axioms and study how violations of an axiom can be fixed. We show that the average error of Bayesian decision makers forms the unique class of utility measures that satisfy all of the axioms. The axioms are agnostic to Bayesian concepts such as subjective probabilities and hence strengthen support for Bayesian views in privacy research. This result connects information preservation to aspect of usability. Utility is measured in Bayesian manner. Shouldn't Bayesian decision theory be a good choice to work with sanitized data? An application to the sorted histogram problem shows that our decision-theoretic post-processing algorithm can estimate the underlying sorted histogram with much greater accuracy than prior work.
590 $a School code: 0176.
650 4 $a Computer Science. $3 626642
650 4 $a Web Studies. $3 1026830
690 $a 0984
690 $a 0646
710 2 $a The Pennsylvania State University. $b Computer Science and Engineering. $3 2095963
773 0 $t Dissertation Abstracts International $g 75-11B(E).
790 $a 0176
791 $a Ph.D.
792 $a 2014
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3583379