There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining - Concepts and Techniques.

In several distributed systems a user should only be able to access data if a user posses a certain set of credentials or attributes. Currently, the only method for enforcing such policies is to employ a trusted server to store the data and mediate access control. However, if any server storing the data is compromised, then the confidentiality of the data will be compromised. In this paper we present a system for realizing complex access control on encrypted data that we call ciphertext-policy attribute-based encryption. By using our techniques encrypted data can be kept confidential even if the storage server is untrusted; moreover, our methods are secure against collusion attacks. Previous attribute-based encryption systems used attributes to describe the encrypted data and built policies into user's keys; while in our system attributes are used to describe a user's credentials, and a party encrypting data determines a policy for who can decrypt. Thus, our methods are conceptually closer to traditional access control methods such as role-based access control (RBAC). In addition, we provide an implementation of our system and give performance measurements.

/pdf/ciphertext-policy-attribute-based-encryption-q09jxdyx4n.pdf

Ciphertext-Policy Attribute-Based Encryption

In 1974 an article appeared in Science magazine with the dry-sounding title “Judgment Under Uncertainty: Heuristics and Biases” by a pair of psychologists who were not well known outside their discipline of decision theory. In it Amos Tversky and Daniel Kahneman introduced the world to Prospect Theory, which mapped out how humans actually behave when faced with decisions about gains and losses, in contrast to how economists assumed that people behave. Prospect Theory turned Economics on its head by demonstrating through a series of ingenious experiments that people are much more concerned with losses than they are with gains, and that framing a choice from one perspective or the other will result in decisions that are exactly the opposite of each other, even if the outcomes are monetarily the same. Prospect Theory led cognitive psychology in a new direction that began to uncover other human biases in thinking that are probably not learned but are part of our brain’s wiring.

Thinking fast and slow.

Data anonymization techniques based on the k-anonymity model have been the focus of intense research in the last few years. Although the k-anonymity model and the related techniques provide valuable solutions to data privacy, current solutions are limited only to static data release (i.e., the entire dataset is assumed to be available at the time of release). While this may be acceptable in some applications, today we see databases continuously growing everyday and even every hour. In such dynamic environments, the current techniques may suffer from poor data quality and/or vulnerability to inference. In this paper, we analyze various inference channels that may exist in multiple anonymized datasets and discuss how to avoid such inferences. We then present an approach to securely anonymizing a continuously growing dataset in an efficient manner while assuring high data quality.

/pdf/secure-anonymization-for-incremental-datasets-4ctx8w63ti.pdf

Secure anonymization for incremental datasets

In recent years, many approaches to differentially privately publish histograms have been proposed. Several approaches rely on constructing tree structures in order to decrease the error when answer large range queries. In this paper, we examine the factors affecting the accuracy of hierarchical approaches by studying the mean squared error (MSE) when answering range queries. We start with one-dimensional histograms, and analyze how the MSE changes with different branching factors, after employing constrained inference, and with different methods to allocate the privacy budget among hierarchy levels. Our analysis and experimental results show that combining the choice of a good branching factor with constrained inference outperform the current state of the art. Finally, we extend our analysis to multi-dimensional histograms. We show that the benefits from employing hierarchical methods beyond a single dimension are significantly diminished, and when there are 3 or more dimensions, it is almost always better to use the Flat method instead of a hierarchy.

/pdf/understanding-hierarchical-methods-for-differentially-4xxadknzdk.pdf

Understanding hierarchical methods for differentially private histograms

The discovery of frequent itemsets can serve valuable economic and research purposes. Releasing discovered frequent itemsets, however, presents privacy challenges. In this paper, we study the problem of how to perform frequent itemset mining on transaction databases while satisfying differential privacy. We propose an approach, called PrivBasis, which leverages a novel notion called basis sets. A θ-basis set has the property that any itemset with frequency higher than θ is a subset of some basis. We introduce algorithms for privately constructing a basis set and then using it to find the most frequent itemsets. Experiments show that our approach greatly outperforms the current state of the art.

/pdf/privbasis-frequent-itemset-mining-with-differential-privacy-3nbl4u58hr.pdf

PrivBasis: frequent itemset mining with differential privacy

In automated trust negotiation (ATN), two parties exchange digitally signed credentials that contain attribute information to establish trust and make access control decisions. Because the information in question is often sensitive, credentials are protected according to access control policies. In traditional ATN, credentials are transmitted either in their entirety or not at all. This approach can at times fail unnecessarily, either because a cyclic dependency makes neither negotiator willing to reveal her credential before her opponent, because the opponent must be authorized for all attributes packaged together in a credential to receive any of them, or because it is necessary to fully disclose exact attribute values, rather than merely proving they satisfy some predicate (such as being over 21 years of age). Recently, several cryptographic credential schemes and associated protocols have been developed to address these and other problems. However, they can be used only as fragments of an ATN process. This paper introduces a framework for ATN in which the diverse credential schemes and protocols can be combined, integrated, and used as needed. A policy language is introduced that enables negotiators to specify authorization requirements that must be met by an opponent to receive various amounts of information about certified attributes and the credentials that contain it. The language also supports the use of uncertified attributes, allowing them to be required as part of policy satisfaction, and to place their (automatic) disclosure under policy control.

/pdf/automated-trust-negotiation-using-cryptographic-credentials-1sxeulggrc.pdf

Automated trust negotiation using cryptographic credentials

The k-anonymity privacy requirement for publishing microdata requires that each equivalence class (i.e., a set of records that are indistinguishable from each other with respect to certain “identifying” attributes) contains at least k records. Recently, several authors have recognized that k-anonymity cannot prevent attribute disclosure. The notion of l-diversity has been proposed to address this; l-diversity requires that each equivalence class has at least l well-represented (in Section 2) values for each sensitive attribute. In this paper, we show that l-diversity has a number of limitations. In particular, it is neither necessary nor sufficient to prevent attribute disclosure. Motivated by these limitations, we propose a new notion of privacy called “closeness.” We first present the base model t-closeness, which requires that the distribution of a sensitive attribute in any equivalence class is close to the distribution of the attribute in the overall table (i.e., the distance between the two distributions should be no more than a threshold t). We then propose a more flexible privacy model called (n,t)-closeness that offers higher utility. We describe our desiderata for designing a distance measure between two probability distributions and present two distance measures. We discuss the rationale for using closeness as a privacy measure and illustrate its advantages through examples and experiments.

Ninghui Li

Papers

Secure anonymization for incremental datasets

Understanding hierarchical methods for differentially private histograms

PrivBasis: frequent itemset mining with differential privacy

Automated trust negotiation using cryptographic credentials

Closeness: A New Privacy Measure for Data Publishing