On September 24-25, 2013, the Privacy Tools for Sharing Research Data project at Harvard University held a workshop titled "Integrating Approaches to Privacy across the Research Data Lifecycle." Over forty leading experts in computer science, statistics, law, policy, and social science research convened to discuss the state of the art in data privacy research. The resulting conversations centered on the emerging tools and approaches from the participants’ various disciplines and how they should be integrated in the context of real-world use cases that involve the management of confidential research data.
Researchers are increasingly obtaining data from social networking websites, publicly-placed sensors, government records and other public sources. Much of this information appears public, at least to first impressions, and it is capable of being used in research for a wide variety of purposes with seemingly minimal legal restrictions. The insights about human behaviors we may gain from research that uses this data are promising. However, members of the research community are questioning the ethics of these practices, and at the heart of the matter are some difficult questions about the boundaries between public and private information. This workshop report, the second in a series, identifies selected questions and explores issues around the meaning of “public” in the context of using data about individuals for research purposes.
We show an essentially tight bound on the number of adaptively chosen statistical queries that a computationally efficient algorithm can answer accurately given n samples from an unknown distribution. A statistical query asks for the expectation of a predicate over the underlying distribution, and an answer to a statistical query is accurate if it is “close” to the correct expectation over the distribution. This question was recently studied by Dwork et al. (2015), who showed how to answer Ω( ˜ n 2 ) queries efficiently, and also by Hardt and Ullman (2014), who showed that answering O˜(n 3 ) queries is hard. We close the gap between the two bounds and show that, under a standard hardness assumption, there is no computationally efficient algorithm that, given n samples from an unknown distribution, can give valid answers to O(n 2 ) adaptively chosen statistical queries. An implication of our results is that computationally efficient algorithms for answering arbitrary, adaptively chosen statistical queries may as well be differentially private. We obtain our results using a new connection between the problem of answering adaptively chosen statistical queries and a combinatorial object called an interactive fingerprinting code Fiat and Tassa (2001). In order to optimize our hardness result, we give a new Fourier-analytic approach to analyzing fingerprinting codes that is simpler, more flexible, and yields better parameters than previous constructions.
A private learner is an algorithm that given a sample of labeled individual examples outputs a generalizing hypothesis while preserving the privacy of each individual. In 2008, Kasiviswanathan et al. (FOCS 2008) gave a generic construction of private learners, in which the sample complexity is (generally) higher than what is needed for non-private learners. This gap in the sample complexity was then further studied in several followup papers, showing that (at least in some cases) this gap is unavoidable. Moreover, those papers considered ways to overcome the gap, by relaxing either the privacy or the learning guarantees of the learner. We suggest an alternative approach, inspired by the (non-private) models of semi-supervised learning and active-learning, where the focus is on the sample complexity of labeled examples whereas unlabeled examples are of a significantly lower cost. We consider private semi-supervised learners that operate on a random sample, where only a (hopefully small) portion of this sample is labeled. The learners have no control over which of the sample elements are labeled. Our main result is that the labeled sample complexity of private learners is characterized by the VC dimension. We present two generic constructions of private semi-supervised learners. The first construction is of learners where the labeled sample complexity is proportional to the VC dimension of the concept class, however, the unlabeled sample complexity of the algorithm is as big as the representation length of domain elements. Our second construction presents a new technique for decreasing the labeled sample complexity of a given private learner, while only slightly increasing its unlabeled sample complexity. In addition, we show that in some settings the labeled sample complexity does not depend on the privacy parameters of the learner.
Imagine an online work environment where researchers have direct and immediate access to myriad data sources and tools and data management resources, useful throughout the research lifecycle. This is our vision for the next generation of the Dataverse Network: an Open Science Platform (OSP). For the first time, researchers would be able to seamlessly access and create primary and derived data from a variety of sources: prior research results, public data sets, harvested online data, physical instruments, private data collections, and even data from other standalone repositories. Researchers could recruit research participants and conduct research directly on the OSP, if desired, using readily available tools. Researchers could create private or shared workspaces to house data, access tools, and computation and could publish data directly on the platform or publish elsewhere with persistent, data citations on the OSP. This manuscript describes the details of an Open Science Platform and its construction. Having an Open Science Platform will especially impact the rate of new scientific discoveries and make scientific findings more credible and accountable.
We introduce three differentially-private algorithms that approximates the 2nd-moment matrix of the data. These algorithm, which in contrast to existing algorithms output positive-definite matrices, correspond to existing techniques in linear regression literature. Specifically, we discuss the following three techniques. (i) For Ridge Regression, we propose setting the regularization coefficient so that by approximating the solution using Johnson-Lindenstrauss transform we preserve privacy. (ii) We show that adding a small batch of random samples to our data preserves differential privacy. (iii) We show that sampling the 2nd-moment matrix from a Bayesian posterior inverse-Wishart distribution is differentially private provided the prior is set correctly. We also evaluate our techniques experimentally and compare them to the existing "Analyze Gauss" algorithm of Dwork et al.
The privacy risks inherent in the release of a large number of summary statistics were illustrated by Homer et al. (PLoS Genetics, 2008), who considered the case of 1-way marginals of SNP allele frequencies obtained in a genome-wide association study: Given a large number of minor allele frequencies from a case group of individuals diagnosed with a particular disease, together with the genomic data of a single target individual and statistics from a sizable reference dataset independently drawn from the same population, an attacker can determine with high confidence whether or not the target is in the case group. In this work we describe and analyze a simple attack that succeeds even if the summary statistics are significantly distorted, whether due to measurement error or noise intentionally introduced to protect privacy. Our attack only requires that the vector of distorted summary statistics is close to the vector of true marginals in `1 norm. Moreover, the reference pool required by previous attacks can be replaced by a single sample drawn from the underlying population. The new attack, which is not specific to genomics and which handles Gaussian as well as Bernouilli data, significantly generalizes recent lower bounds on the noise needed to ensure differential privacy (Bun, Ullman, and Vadhan, STOC 2014; Steinke and Ullman, 2015), obviating the need for the attacker to control the exact distribution of the data.
Society generates data on a scale previously unimagined. Wide sharing of these data promises to improve personal health, lower healthcare costs, and provide a better quality of life. There is a tendency to want to share data freely. However, these same data often include sensitive information about people that could cause serious harms if shared widely. A multitude of regulations, laws and best practices protect data that contain sensitive personal information. Government agencies, research labs, and corporations that share data, as well as review boards and privacy officers making data sharing decisions, are vigilant but uncertain. This uncertainty creates a tendency not to share data at all. Some data are more harmful than other data; sharing should not be an all-or-nothing choice. How do we share data in ways that ensure access is commensurate with risks of harm?
We investigate the direct-sum problem in the context of differentially private PAC learning: What is the sample complexity of solving k learning tasks simultaneously under differential privacy, and how does this cost compare to that of solving k learning tasks without privacy? In our setting, an individual example consists of a domain element x labeled by k unknown concepts (c1,…,ck). The goal of a multi-learner is to output k hypotheses (h1,…,hk) that generalize the input examples. Without concern for privacy, the sample complexity needed to simultaneously learn k concepts is essentially the same as needed for learning a single concept. Under differential privacy, the basic strategy of learning each hypothesis independently yields sample complexity that grows polynomially with k. For some concept classes, we give multi-learners that require fewer samples than the basic strategy. Unfortunately, however, we also give lower bounds showing that even for very simple concept classes, the sample cost of private multi-learning must grow polynomially in k.
On January 23, 2014, President Barack Obama asked John Podesta to perform a comprehensive review of big data and privacy. During this review, the White House Office of Science and Technology Policy issued a request for public comment on questions related to the public policy implications of big data.
Micah Altman, David O’Brien, Salil Vadhan, and Alexandra Wood submitted a response on behalf of the Privacy Tools for Sharing Research Data project. Their comments outline a broad, comprehensive, and systematic framework for privacy analysis and provide a taxonomy of modern technological, statistical, and cryptographic approaches to preserving both data privacy and utility. They argue that an analysis of information privacy should address the scope of information covered, the sensitivity of that information, the risk that sensitive information will be disclosed, the availability of control and accountability mechanisms, and the suitability of existing data sharing models, as applied across the entire lifecyle of information use, from collection through dissemination and reuse.
With this submission, the authors discuss the inadequacy of traditional approaches to privacy protection and recommend a modern approach that considers three principles. First, the risks of informational harm are generally not a simple function of the presence or absence of specific fields, attributes, or keywords in the released set of data. Second, redaction, pseudonymization, coarsening, and hashing, are often neither an adequate nor appropriate practice, nor is releasing less information necessary more privacy protective. Third, a thoughtful analysis with expert consultation is necessary in order to evaluate the sensitivity of the data collected, to quantify the associated re-identification risks, and to design useful and safe release mechanisms.
In response to the White House Office of Science and Technology Policy Request for Information on Big Data Privacy we offer these comments based on presentations and discussions at the White House-MIT Workshop “Big Data Privacy Workshop: Advancing the State of the Art in Technology and Practice” and subsequent workshops co-sponsored with Data & Society and NYU Information Law Institute and the UC Berkeley iSchool.
In capability-safe languages, components can access a resource only if they possess a capability for that resource. As a result, a programmer can prevent an untrusted component from accessing a sensitive resource by ensuring that the component never acquires the corresponding capability. In order to reason about which components may use a sensitive resource it is necessary to reason about how capabilities propagate through a system. This may be difficult, or, in the case of dynamically composed code, impossible to do before running the system.
To counter this situation, we propose extensions to capability-safe languages that restrict the use of capabilities according to declarative policies. We introduce two independently useful semantic security policies to regulate capabilities and describe language-based mechanisms that enforce them. Access control policies restrict which components may use a capability and are enforced using higher-order contracts. Integrity policies restrict which components may influence (directly or indirectly) the use of a capability and are enforced using an information-flow type system. Finally, we describe how programmers can dynamically and soundly combine components that enforce access control or integrity policies with components that enforce different policies or even no policy at all.