Privacy definitions provide ways for trading-off the privacy of individuals in a statistical database utility downstream analysis data. In this paper, we present Blowfish, class inspired by Pufferfish framework, that provides rich interface trade-off. particular, allow data publishers to extend differential using policy, which specifies (a) secrets, or information must be kept secret, and (b) constraints may known about While secret specification allows increased lessening protection certain...

Private record linkage (PRL) is the problem of identifying pairs records that are similar as per an input matching rule from databases held by two parties do not trust one another. We identify three key desiderata a PRL solution must ensure: (1) perfect precision and high recall pairs, (2) proof end-to-end privacy, (3) communication computational costs scale subquadratically in number records. show all existing solutions for PRL? including secure 2-party computation (S2PC), their variants...

Abstract Organizations often collect private data and release aggregate statistics for the public’s benefit. If no steps toward preserving privacy are taken, adversaries may use released to deduce unauthorized information about individuals described in dataset. Differentially algorithms address this challenge by slightly perturbing underlying with noise, thereby mathematically limiting amount of that be deduced from each release. Properly calibrating these algorithms—and turn disclosure risk...

With the severe plastic pollution issue worldwide, restrictions or bans on bags have become most popular policy intervention. As essential participants in reducing consumption, residents and enterprises are vital implementing bag restriction law (PBRL). Through a questionnaires survey of 630 consumers 50 enterprises, this study investigates residents’ enterprises’ perceptions behavioural changes toward PBRL Macao identifies key influence factors. The results show that respondents (95%) began...

A private data federation enables clients to query the union of from multiple providers without revealing any extra information client or other providers. Unfortunately, this strong end-to-end privacy guarantee requires cryptographic protocols that incur a significant performance overhead as high 1,000 x compared executing same in clear. As result, federations are impractical for common database workloads. This gap reveals following key challenge federation: offering significantly fast and...

Local sensitivity of a query Q given database instance D, i.e. how much the output Q(D) changes when tuple is added to D or deleted from has many applications including analysis, outlier detection, and in differential privacy. However, it NP-hard find local conjunctive terms size query, even for class acyclic queries. Although complexity polynomial fixed, naive algorithms are not efficient large databases queries involving multiple joins. In this paper, we present novel approach compute...

We present a method for producing unbiased parameter estimates and valid confidence regions/intervals under the constraints of differential privacy, formal framework limiting individual information leakage from sensitive data.Prior work in this area is limited that it tailored to calculating intervals specific statistical procedures, such as mean estimation or simple linear regression.While other recent can produce more general sets they either yield only approximately estimates, are...

An efficient parallel priority queue is at the core of effort in parallelizing important non-numeric irregular computations such as discrete event simulation scheduling and branch-and-bound algorithms. GPGPUs can provide powerful computing platform for if an implementation available. In this paper, aiming fine-grained applications, we develop heap system employing CUDA. To our knowledge, first on many-core architectures, thus represents a breakthrough. By allowing wide nodes to enable...

Sports data visualization can be a useful tool for analyzing or presenting sports data. In this paper, we present new technique visualizing tennis match It is designed as supplement to online live streaming blogging of matches. retrieve directly from web site and display 2D interactive view statistics. Therefore, it easily integrated with the current platforms used by many news organizations. The addresses limitations coverage matches providing quick overview also great amount details on...

Hyperparameter optimization is a ubiquitous challenge in machine learning, and the performance of trained model depends crucially upon their effective selection. While rich set tools exist for this purpose, there are currently no practical hyperparameter selection methods under constraint differential privacy (DP). We study honest differentially private which process tuning accounted overall budget. To end, we i) show that standard composition outperform more advanced techniques many...

A private data federation is a set of autonomous databases that share unified query interface offering in-situ evaluation SQL queries over the union sensitive its members. Owing to privacy concerns, these systems do not have trusted collector can see all their and member cannot learn about individual records other engines. Federations currently achieve this goal by evaluating obliviously using secure multiparty computation. This hides intermediate result cardinality each operator...

Differential privacy (DP) allows data analysts to query databases that contain users' sensitive information while providing a quantifiable guarantee users. Recent interactive DP systems such as APEx provide accuracy guarantees over the responses, but fail support large number of queries with limited total budget, they process incoming independently from past queries. We present an interactive, accuracy-aware engine, CacheDP , which utilizes differentially private cache answer current...

Organizations are increasingly relying on data to support decisions. When contains private and sensitive information, the owner often desires publish a synthetic database instance that is similarly useful as true data, while ensuring privacy of individual records. Existing differentially synthesis methods aim generate based applications, but they fail in keeping one most fundamental properties structured -- underlying correlations dependencies among tuples attributes (i.e., structure data)....

Differential privacy has become an appealing choice for analyzing sensitive data while offering strong protection, even complex types like graphs. Despite a decade of academic efforts in designing differentially private algorithms graph analysis, few works have been used practice. This is due to their complexity the guarantees and parameter/environmental configurations, or scalability issues large datasets.

Particle Swarm Optimization (PSO) has attracted many researchers attention to solve variant benchmark and real-world optimization problems because of its simplicity, effective performance fast convergence. However, it suffers from premature convergence quickly losing diversity. To enhance performance, this paper proposes a novel disruption strategy, originating astrophysics, shift the abilities between exploration exploitation. The proposed Disruption PSO (DPSO) been evaluated on set...