Tentatively, there will be no lectures on these dates: January 27; February 3, 8, 15, and 17; March 8, 10, 15, 24, 29, and 31; and April 12 and 14.
- Title: Legal Compliance, Privacy and Security
- Abstract: U.S. federal and state regulations impose mandatory and discretionary requirements on industry-wide business practices to achieve non-functional, societal goals such as improved privacy and security. The structure and syntax of regulations affects how well software engineers identify and interpret legal requirements. Inconsistent interpretations can lead to non-compliance and violations of privacy law. To support software engineers who must comply with these regulations, I discuss several research challenges and recent results in the form of validated methods and tools to systematically identify privacy requirements from laws and regulations. These results are motivated by examples from the Health Insurance Portability and Accountability Act of 1996.
Philosophy and Law meets Computer Science
Foundations of Privacy
- Title: The Economics and Behavioral Economics of Privacy
|Talks and papers
- Bio: Prof. Ramayya Krishnan is Dean of the Heinz College and Director of its
iLab. His current interests are in social media analytics and in information
privacy and risk management. He can be reached at email@example.com.
- Title: On Protecting Privacy in Published Network Data
- Abstract: Social network data is increasingly ubiquitous. Publishing network data
while protecting privacy is a challenging problem. Extant approaches attempt
to add "noise" to the original network to protect privacy. While they
attempt to preserve properties of the original network, extant approaches
have not taken the needs of network statistical analysis into account.
Using ongoing work on network analysis which does not focus on privacy as a
point of departure, I will highlight the problems that arise in conducting
network statistical analysis when approaches to protect privacy are applied.
The talk will also highlight the availability of large, societal scale
network data sets available at the Heinz College iLab and their potential to
support research in privacy and in social network analysis.
|Hay, Miklau, and Jensen chapter
- Title: An Overview for Location Privacy for Mobile Computing
- .ppt slides
|HCI and privacy survey paper
- Title: A brief tour of differential privacy
- pptx, pdf
Kobbi Nissim, Sofya Raskhodnikova and Adam Smith: Smooth Sensitivity
and Sampling in Private Data Analysis
Authors: Kamalika Chaudhuri, Claire Monteleoni, Anand D. Sarwate:
Differentially Private Empirical Risk Minimization
Avrim Blum, Katrina Ligett and Aaron Roth: A Learning Theory Approach
to Non-Interactive Database Privacy
- Larry Wasserman is a Professor in the Department of Statistics and the Machine Learning Department.
- Title: A Statistical Framework for Differential Privacy
I'll review differential privacy and discuss how differential privacy
affects the accuracy of some statistical procedures. I'll also
discuss some shortcomings of differential privacy. (Joint work with
- Norman Sadeh is a Professor in the School of Computer Science at Carnegie Mellon
University. His current research interests include Web Security, Privacy and Commerce.
- Title: User-Controllable Privacy: A Multi-Disciplinary Perspective
- Abstract: Increasingly users are expected to evaluate and configure a variety of privacy policies (e.g. browser settings, mobile app manifests, or social networking accounts). In practice, research shows that users often have great difficulty evaluating and configuring such policies. As part of this presentation, I will provide an overview of research aimed at empowering users to better control their privacy in the context of a family of location sharing applications we have deployed over the years. This includes technologies to analyze people.s privacy preferences and help design interfaces that are capable of effectively capturing their desired policies. This research helps explain why, with the possible exception of Foursquare, applications in this space have failed to gain traction and what it will likely take to go beyond the mundane scenarios captured by Foursquare. A good part of this talk will be devoted to user-oriented machine learning techniques intended to reduce user-burden and help users converge towards policies they feel more comfortable with. I will also discuss how, beyond just capturing people.s preferences, these same techniques could possibly be used to nudge users towards safer privacy practices.
- Marco Gruteser is an Associate Professor of Electrical and Computer
Engineering at Rutgers University. He is a visiting CMU this year, working with Lorrie
- Title: Wireless Location Privacy: Depersonalization Techniques and Connected Vehicle Applications
- Stephen Fienberg is the Maurice Falk University Professor of Statistics and Social Science in the Department of Statistics, the Machine Learning Department, CyLab, and i-Lab.
- Title: Statistical Disclosure Limitation and the Challenge of Societal-Scale Data.
- Srini Seshan is an Associate Professor of Computer Science in the Computer Science Department.
- Title: Improving the Privacy of Wireless Protocols
- Bhiksha Raj is an Associate Professor in the Lnaguage Technologies Institute.
- Title: Privacy issues in speech processing.
- Abstract: Speech is perhaps one of the most private forms of communication. A
person's speech conveys not only what the person says, but also their
identity, their emotional state and other such information that the
speaker may not want revealed to anyone besides their intended
audience. Legally too, the privacy of speech has been recognized: in
fact, in many places it is considered illegal to record a person's
voice in public even when it is legal to capture their images in
Yet, in spite of the significant theoretical and practical advances in
privacy and security technology, little of it has been applied to
In this talk we will present the privacy issues related to various
typical voice applications. We will also briefly describe some of our
current research directions, and some of the basic tools currently
available to develop solutions.
- John Lafferty is a Professor of Computer Science, Machine Learning, and Statistics.
- Title: Compressed Regression
We present results on a variant of the classical linear regression
problem where the original input records are compressed by a random
linear transformation. A primary motivation for this compression
procedure is to anonymize the data and preserve privacy by revealing
only weighted linear combinations of the original observations. We
characterize the number of random projections that are required for
l1-regularized compressed regression to identify the nonzero
coefficients in the true model with probability approaching one. In
addition, we show that l1-regularized compressed regression
asymptotically predicts as well as an oracle linear model. Finally,
we characterize the privacy properties of the compression procedure in
information-theoretic terms, establishing upper bounds on the mutual
information between the compressed and uncompressed data that decay to
Joint work with Larry Wasserman (CMU) and Shuheng Zhou (University of
Michigan). (Appeared in IEEE Trans. Info. Theory, Vol 55, No. 2, 2009.)