Ninghui Li's guest appearances

Ninghui Li

CERIAS Weekly Security Seminar - Purdue University

Ryan Henry

Cory Doctorow

Doug Smith

Larry Ponemon

George Heron

Laura Thomas

Eli Lilly

Stephen Elliott

Tony Sager

David Carroll

Jack Daniel

Joe Judge

Chris Reed

Richard III

John Richardson

OwlTail

Data collected by organizations and agencies are a key resource intoday's information age.  The use of sophisticated data mining techniquesmakes it possible to extract relevant knowledge that can then be used for avariety of purposes, such as research, developing innovative technologiesand services, intelligence and counterterrorism operations, and providinginputs to public policy making.  However the disclosure of those data posesserious threats to individual privacy.       In this talk, we will present the evolvement of privacy notions fordata publishing and analysis, leading to our proposed membership privacyframework, which formalizes the intuition that privacy means that theadversary cannot significantly increasing its ability to conclude that anentity is in the input dataset.  We show that several recently proposedprivacy notions, including differential privacy, are instantiations of themembership privacy framework, and that the framework provides a principledapproach to developing new privacy notions under which better utility can beachieved than what is possible under differential privacy. About the speaker: Ninghui Li is a Professor of Computer Science at Purdue University.  Hisresearch interests are insecurity and privacy.   Prof. Li is currently Vice Chair of ACM SpecialInterest Group on Security, Audit and Control (SIGSAC) and Program Chair of 2015 ACM Conference on Computer andCommunications Security (CCS).  He is on the editorial boards of IEEE Transactions on Dependable and SecureComputing, Journal of Computer Security, and ACM Transactions on InternetTechnology.

<p>Data collected by organizations and agencies are a key resource intoday's information age.  The use of sophisticated data mining techniquesmakes it possible to extract relevant knowledge that can then be used for avariety of purposes, such as research, developing innovative technologiesand services, intelligence and counterterrorism operations, and providinginputs to public policy making.  However the disclosure of those data posesserious threats to individual privacy.       In this talk, we will present the evolvement of privacy notions fordata publishing and analysis, leading to our proposed membership privacyframework, which formalizes the intuition that privacy means that theadversary cannot significantly increasing its ability to conclude that anentity is in the input dataset.  We show that several recently proposedprivacy notions, including differential privacy, are instantiations of themembership privacy framework, and that the framework provides a principledapproach to developing new privacy notions under which better utility can beachieved than what is possible under differential privacy. About the speaker: Ninghui Li is a Professor of Computer Science at Purdue University.  Hisresearch interests are insecurity and privacy.   Prof. Li is currently Vice Chair of ACM SpecialInterest Group on Security, Audit and Control (SIGSAC) and Program Chair of 2015 ACM Conference on Computer andCommunications Security (CCS).  He is on the editorial boards of IEEE Transactions on Dependable and SecureComputing, Journal of Computer Security, and ACM Transactions on InternetTechnology.</p>

Ninghui Li, Privacy Notions for Data Publishing and Analysis

Data collected by organizations and agencies are a key resourcein today's information age. The use of sophisticated data mining techniquesmakes it possible to extract relevant knowledge that can then be used for avariety of purposes, such as research, developing innovative technologiesand services, intelligence and counter-terrorism operations, and providinginputs to public policy making. However the disclosure of those data posesserious threats to individual privacy.  In this talk, we present a novelprivacy framework that we call Membership Privacy, which prevents theadversary from significantly increasing its ability to conclude that anentity is in the input dataset.  Membership privacy is parameterized by afamily of distributions that captures the adversary's prior knowledge. Thepower and flexibility of the proposed framework lies in the ability tochoose different distribution families to instantiate membership privacy.Many privacy notions in the literature are equivalent to membership privacywith interesting distribution families, including differential privacy,differential identifiability, and differential privacy under sampling.  Theframework also provides a principled approach to developing new privacynotions under which better utility can be achieved than what is possibleunder differential privacy.  This is joint work with Wahbeh Qardaji, DongSu, Yi Wu, and Weining Yang. About the speaker: Ninghui Li is an Associate Professor of Computer Science atPurdue University.  He received a Bachelor's degree from the University ofScience and Technology of China in 1993 and a Ph.D. in Computer Science fromNew York University in 2000. Before joining the faculty of Purdue in 2003,he was a Research Associate at Stanford University Computer ScienceDepartment for 3 years.   Prof. Li's research interests are in security andprivacy, and has published over 100 referred papers in this area.  Prof. Liis an ACM Distinguished Scientist and IEEE Senior member.  In June 2013, hewas elected Vice Chair of ACM Special Interest Group on Security, Audit andControl (SIGSAC).  He served on the editorial board of the VLDB Journal from2007 to 2013, and is current on the editorial board of IEEE Transactions onDependable and Secure Computing (TDSC).

<p>Data collected by organizations and agencies are a key resourcein today's information age. The use of sophisticated data mining techniquesmakes it possible to extract relevant knowledge that can then be used for avariety of purposes, such as research, developing innovative technologiesand services, intelligence and counter-terrorism operations, and providinginputs to public policy making. However the disclosure of those data posesserious threats to individual privacy.  In this talk, we present a novelprivacy framework that we call Membership Privacy, which prevents theadversary from significantly increasing its ability to conclude that anentity is in the input dataset.  Membership privacy is parameterized by afamily of distributions that captures the adversary's prior knowledge. Thepower and flexibility of the proposed framework lies in the ability tochoose different distribution families to instantiate membership privacy.Many privacy notions in the literature are equivalent to membership privacywith interesting distribution families, including differential privacy,differential identifiability, and differential privacy under sampling.  Theframework also provides a principled approach to developing new privacynotions under which better utility can be achieved than what is possibleunder differential privacy.  This is joint work with Wahbeh Qardaji, DongSu, Yi Wu, and Weining Yang. About the speaker: Ninghui Li is an Associate Professor of Computer Science atPurdue University.  He received a Bachelor's degree from the University ofScience and Technology of China in 1993 and a Ph.D. in Computer Science fromNew York University in 2000. Before joining the faculty of Purdue in 2003,he was a Research Associate at Stanford University Computer ScienceDepartment for 3 years.   Prof. Li's research interests are in security andprivacy, and has published over 100 referred papers in this area.  Prof. Liis an ACM Distinguished Scientist and IEEE Senior member.  In June 2013, hewas elected Vice Chair of ACM Special Interest Group on Security, Audit andControl (SIGSAC).  He served on the editorial board of the VLDB Journal from2007 to 2013, and is current on the editorial board of IEEE Transactions onDependable and Secure Computing (TDSC).</p>

Ninghui Li, Membership Privacy: A Unifying Framework For Privacy Definitions