arXiv subjects

Divesh Srivastava

Publications and source records attributed to Divesh Srivastava.

At least 37 records · Page 2Linked to original sources

arXiv2021

Discovering Domain Orders through Order Dependencies

Much real-world data come with explicitly defined domain orders; e.g., lexicographic order for strings, numeric for integers, and chronological for time. Our goal is to discover implicit domain orders that we do not already know; for instance, that the order of months in the Chinese Lunar calendar is Corner < Apricot < Peach. To do so, we enhance data profiling methods by discovering implicit domain orders in data through order dependencies. We enumerate tractable special cases and proceed towards the most general case, which we prove is NP-complete. We show that the general case nevertheless can be effectively handled by a SAT solver. We also devise an interestingness measure to rank the discovered implicit domain orders, which we validate with a user study. Based on an extensive suite of experiments with real-world data, we establish the efficacy of our algorithms, and the utility of the domain orders discovered by demonstrating significant added value in three applications (data profiling, query optimization, and data mining).

cs.DB↗

arXiv2021

Real-World Trajectory Sharing with Local Differential Privacy

Sharing trajectories is beneficial for many real-world applications, such as managing disease spread through contact tracing and tailoring public services to a population's travel patterns. However, public concern over privacy and data protection has limited the extent to which this data is shared. Local differential privacy enables data sharing in which users share a perturbed version of their data, but existing mechanisms fail to incorporate user-independent public knowledge (e.g., business locations and opening times, public transport schedules, geo-located tweets). This limitation makes mechanisms too restrictive, gives unrealistic outputs, and ultimately leads to low practical utility. To address these concerns, we propose a local differentially private mechanism that is based on perturbing hierarchically-structured, overlapping $n$-grams (i.e., contiguous subsequences of length $n$) of trajectory data. Our mechanism uses a multi-dimensional hierarchy over publicly available external knowledge of real-world places of interest to improve the realism and utility of the perturbed, shared trajectories. Importantly, including real-world public data does not negatively affect privacy or efficiency. Our experiments, using real-world data and a range of queries, each with real-world application analogues, demonstrate the superiority of our approach over a range of alternative methods.

cs.DB↗

arXiv2021

DataExposer: Exposing Disconnect between Data and Systems

As data is a central component of many modern systems, the cause of a system malfunction may reside in the data, and, specifically, particular properties of the data. For example, a health-monitoring system that is designed under the assumption that weight is reported in imperial units (lbs) will malfunction when encountering weight reported in metric units (kilograms). Similar to software debugging, which aims to find bugs in the mechanism (source code or runtime conditions), our goal is to debug the data to identify potential sources of disconnect between the assumptions about the data and the systems that operate on that data. Specifically, we seek which properties of the data cause a data-driven system to malfunction. We propose DataExposer, a framework to identify data properties, called profiles, that are the root causes of performance degradation or failure of a system that operates on the data. Such identification is necessary to repair the system and resolve the disconnect between data and system. Our technique is based on causal reasoning through interventions: when a system malfunctions for a dataset, DataExposer alters the data profiles and observes changes in the system's behavior due to the alteration. Unlike statistical observational analysis that reports mere correlations, DataExposer reports causally verified root causes, in terms of data profiles, of the system malfunction. We empirically evaluate DataExposer on three real-world and several synthetic data-driven systems that fail on datasets due to a diverse set of reasons. In all cases, DataExposer identifies the root causes precisely while requiring orders of magnitude fewer interventions than prior techniques.

cs.DB↗

arXiv2021

Efficient and Effective ER with Progressive Blocking

Blocking is a mechanism to improve the efficiency of Entity Resolution (ER) which aims to quickly prune out all non-matching record pairs. However, depending on the distributions of entity cluster sizes, existing techniques can be either (a) too aggressive, such that they help scale but can adversely affect the ER effectiveness, or (b) too permissive, potentially harming ER efficiency. In this paper, we propose a new methodology of progressive blocking (pBlocking) to enable both efficient and effective ER, which works seamlessly across different entity cluster size distributions. pBlocking is based on the insight that the effectiveness-efficiency trade-off is revealed only when the output of ER starts to be available. Hence, pBlocking leverages partial ER output in a feedback loop to refine the blocking result in a data-driven fashion. Specifically, we bootstrap pBlocking with traditional blocking methods and progressively improve the building and scoring of blocks until we get the desired trade-off, leveraging a limited amount of ER results as a guidance at every round. We formally prove that pBlocking converges efficiently ($O(n log^2 n)$ time complexity, where n is the total number of records). Our experiments show that incorporating partial ER output in a feedback loop can improve the efficiency and effectiveness of blocking by 5x and 60% respectively, improving the overall F-score of the entire ER process up to 60%.

cs.DB↗

arXiv2021

Alaska: A Flexible Benchmark for Data Integration Tasks

Data integration is a long-standing interest of the data management community and has many disparate applications, including business, science and government. We have recently witnessed impressive results in specific data integration tasks, such as Entity Resolution, thanks to the increasing availability of benchmarks. A limitation of such benchmarks is that they typically come with their own task definition and it can be difficult to leverage them for complex integration pipelines. As a result, evaluating end-to-end pipelines for the entire data integration process is still an elusive goal. In this work, we present Alaska, the first benchmark based on real-world dataset to support seamlessly multiple tasks (and their variants) of the data integration pipeline. The dataset consists of ~70k heterogeneous product specifications from 71 e-commerce websites with thousands of different product attributes. Our benchmark comes with profiling meta-data, a set of pre-defined use cases with diverse characteristics, and an extensive manually curated ground truth. We demonstrate the flexibility of our benchmark by focusing on several variants of two crucial data integration tasks, Schema Matching and Entity Resolution. Our experiments show that our benchmark enables the evaluation of a variety of methods that previously were difficult to compare, and can foster the design of more holistic data integration solutions.

cs.DB↗

arXiv2021

Efficient Discovery of Approximate Order Dependencies

Order dependencies (ODs) capture relationships between ordered domains of attributes. Approximate ODs (AODs) capture such relationships even when there exist exceptions in the data. During automated discovery of ODs, validation is the process of verifying whether an OD holds. We present an algorithm for validating approximate ODs with significantly improved runtime performance over existing methods for AODs, and prove that it is correct and has optimal runtime. By replacing the validation step in a leading algorithm for approximate OD discovery with ours, we achieve orders-of-magnitude improvements in performance.

cs.DB↗

arXiv2020

Effective Discovery of Meaningful Outlier Relationships

We propose PODS (Predictable Outliers in Data-trendS), a method that, given a collection of temporal data sets, derives data-driven explanations for outliers by identifying meaningful relationships between them. First, we formalize the notion of meaningfulness, which so far has been informally framed in terms of explainability. Next, since outliers are rare and it is difficult to determine whether their relationships are meaningful, we develop a new criterion that does so by checking if these relationships could have been predicted from non-outliers, i.e., if we could see the outlier relationships coming. Finally, searching for meaningful outlier relationships between every pair of data sets in a large data collection is computationally infeasible. To address that, we propose an indexing strategy that prunes irrelevant comparisons across data sets, making the approach scalable. We present the results of an experimental evaluation using real data sets and different baselines, which demonstrates the effectiveness, robustness, and scalability of our approach.

cs.DB↗

arXiv2020

ABC of Order Dependencies

We enhance constrained-based data quality with approximate band conditional order dependencies (abcODs). Band ODs model the semantics of attributes that are monotonically related with small variations without there being an intrinsic violation of semantics. The class of abcODs generalizes band ODs to make them more relevant to real-world applications by relaxing them to hold approximately (abODs) with some exceptions and conditionally (bcODs) on subsets of the data. We study the problem of automatic dependency discovery over a hierarchy of abcODs. First, we propose a more efficient algorithm to discover abODs than in recent prior work. The algorithm is based on a new optimization to compute a longest monotonic band (longest subsequence of tuples that satisfy a band OD) through dynamic programming by decreasing the runtime from O(n^2) to O(n \log n) time. We then illustrate that while the discovery of bcODs is relatively straightforward, there exist codependencies between approximation and conditioning that make the problem of abcOD discovery challenging. The naive solution is prohibitively expensive as it considers all possible segmentations of tuples resulting in exponential time complexity. To reduce the search space, we devise a dynamic programming algorithm for abcOD discovery that determines the optimal solution in O(n^3 \log n) complexity. To further optimize the performance, we adapt the algorithm to cheaply identify consecutive tuples that are guaranteed to belong to the same band--this improves the performance significantly in practice without losing optimality. While unidirectional abcODs are most common in practice, for generality we extend our algorithms with both ascending and descending orders to discover bidirectional abcODs. Finally, we perform a thorough experimental evaluation of our techniques over real-world and synthetic datasets.

cs.DB↗

arXiv2019

Random Sampling for Group-By Queries

Random sampling has been widely used in approximate query processing on large databases, due to its potential to significantly reduce resource usage and response times, at the cost of a small approximation error. We consider random sampling for answering the ubiquitous class of group-by queries, which first group data according to one or more attributes, and then aggregate within each group after filtering through a predicate. The challenge with group-by queries is that a sampling method cannot focus on optimizing the quality of a single answer (e.g. the mean of selected data), but must simultaneously optimize the quality of a set of answers (one per group). We present CVOPT, a query- and data-driven sampling framework for a set of group-by queries. To evaluate the quality of a sample, CVOPT defines a metric based on the norm (e.g. $\ell_2$ or $\ell_\infty$) of the coefficients of variation (CVs) of different answers, and constructs a stratified sample that provably optimizes the metric. CVOPT can handle group-by queries on data where groups have vastly different statistical characteristics, such as frequencies, means, or variances. CVOPT jointly optimizes for multiple aggregations and multiple group-by clauses, and provides a way to prioritize specific groups or aggregates. It can be tuned to cases when partial information about a query workload is known, such as a data warehouse where queries are scheduled to run periodically. Our experimental results show that CVOPT outperforms current state-of-the-art on sample quality and estimation accuracy for group-by queries. On a set of queries on two real-world data sets, CVOPT yields relative errors that are 5x smaller than competing approaches, under the same space budget.

cs.DB↗

arXiv2019

Errata Note: Discovering Order Dependencies through Order Compatibility

A number of extensions to the classical notion of functional dependencies have been proposed to express and enforce application semantics. One of these extensions is that of order dependencies (ODs), which express rules involving order. The article entitled "Discovering Order Dependencies through Order Compatibility" by Consonni et al., published in the EDBT conference proceedings in March 2019, investigates the OD discovery problem. They claim to prove that their OD discovery algorithm, OCDDISCOVER, is complete, as well as being significantly more efficient in practice than the state-of-the-art. They further claim that the implementation of the existing FASTOD algorithm (ours)-we shared our code base with the authors-which they benchmark against is flawed, as OCDDISCOVER and FASTOD report different sets of ODs over the same data sets. In this rebuttal, we show that their claim of completeness is, in fact, not true. Built upon their incorrect claim, OCDDISCOVER's pruning rules are overly aggressive, and prune parts of the search space that contain legitimate ODs. This is the reason their approach appears to be "faster" in practice. Finally, we show that Consonni et al. misinterpret our set-based canonical form for ODs, leading to an incorrect claim that our FASTOD implementation has an error.

cs.DB↗

arXiv2018

Answering Range Queries Under Local Differential Privacy

Counting the fraction of a population having an input within a specified interval i.e. a \emph{range query}, is a fundamental data analysis primitive. Range queries can also be used to compute other interesting statistics such as \emph{quantiles}, and to build prediction models. However, frequently the data is subject to privacy concerns when it is drawn from individuals, and relates for example to their financial, health, religious or political status. In this paper, we introduce and analyze methods to support range queries under the local variant of differential privacy, an emerging standard for privacy-preserving data analysis. The local model requires that each user releases a noisy view of her private data under a privacy guarantee. While many works address the problem of range queries in the trusted aggregator setting, this problem has not been addressed specifically under untrusted aggregation (local DP) model even though many primitives have been developed recently for estimating a discrete distribution. We describe and analyze two classes of approaches for range queries, based on hierarchical histograms and the Haar wavelet transform. We show that both have strong theoretical accuracy guarantees on variance. In practice, both methods are fast and require minimal computation and communication resources. Our experiments show that the wavelet approach is most accurate in high privacy settings, while the hierarchical approach dominates for weaker privacy requirements.

cs.DB↗

arXiv2018

Variance-Optimal Offline and Streaming Stratified Random Sampling

Stratified random sampling (SRS) is a fundamental sampling technique that provides accurate estimates for aggregate queries using a small size sample, and has been used widely for approximate query processing. A key question in SRS is how to partition a target sample size among different strata. While Neyman allocation provides a solution that minimizes the variance of an estimate using this sample, it works under the assumption that each stratum is abundant, i.e., has a large number of data points to choose from. This assumption may not hold in general: one or more strata may be bounded, and may not contain a large number of data points, even though the total data size may be large. We first present VOILA, an offline method for allocating sample sizes to strata in a variance-optimal manner, even for the case when one or more strata may be bounded. We next consider SRS on streaming data that are continuously arriving. We show a lower bound, that any streaming algorithm for SRS must have (in the worst case) a variance that is Ω(r) factor away from the optimal, where r is the number of strata. We present S-VOILA, a practical streaming algorithm for SRS that is locally variance-optimal in its allocation of sample sizes to different strata. Our result from experiments on real and synthetic data show that VOILA can have significantly (1.4 to 50.0 times) smaller variance than Neyman allocation. The streaming algorithm S-VOILA results in a variance that is typically close to VOILA, which was given the entire input beforehand.

cs.DB↗

arXiv2017

Marginal Release Under Local Differential Privacy

Many analysis and machine learning tasks require the availability of marginal statistics on multidimensional datasets while providing strong privacy guarantees for the data subjects. Applications for these statistics range from finding correlations in the data to fitting sophisticated prediction models. In this paper, we provide a set of algorithms for materializing marginal statistics under the strong model of local differential privacy. We prove the first tight theoretical bounds on the accuracy of marginals compiled under each approach, perform empirical evaluation to confirm these bounds, and evaluate them for tasks such as modeling and correlation testing. Our results show that releasing information based on (local) Fourier transformations of the input is preferable to alternatives based directly on (local) marginals.

cs.DB↗

arXiv2017

Constrained Differential Privacy for Count Data

Concern about how to aggregate sensitive user data without compromising individual privacy is a major barrier to greater availability of data. The model of differential privacy has emerged as an accepted model to release sensitive information while giving a statistical guarantee for privacy. Many different algorithms are possible to address different target functions. We focus on the core problem of count queries, and seek to design mechanisms to release data associated with a group of n individuals. Prior work has focused on designing mechanisms by raw optimization of a loss function, without regard to the consequences on the results. This can leads to mechanisms with undesirable properties, such as never reporting some outputs (gaps), and overreporting others (spikes). We tame these pathological behaviors by introducing a set of desirable properties that mechanisms can obey. Any combination of these can be satisfied by solving a linear program (LP) which minimizes a cost function, with constraints enforcing the properties. We focus on a particular cost function, and provide explicit constructions that are optimal for certain combinations of properties, and show a closed form for their cost. In the end, there are only a handful of distinct optimal mechanisms to choose between: one is the well-known (truncated) geometric mechanism; the second a novel mechanism that we introduce here, and the remainder are found as the solution to particular LPs. These all avoid the bad behaviors we identify. We demonstrate in a set of experiments on real and synthetic data which is preferable in practice, for different combinations of data distributions, constraints, and privacy parameters.

cs.DB↗

arXiv2017

Composing Differential Privacy and Secure Computation: A case study on scaling private record linkage

Private record linkage (PRL) is the problem of identifying pairs of records that are similar as per an input matching rule from databases held by two parties that do not trust one another. We identify three key desiderata that a PRL solution must ensure: 1) perfect precision and high recall of matching pairs, 2) a proof of end-to-end privacy, and 3) communication and computational costs that scale subquadratically in the number of input records. We show that all of the existing solutions for PRL - including secure 2-party computation (S2PC), and their variants that use non-private or differentially private (DP) blocking to ensure subquadratic cost - violate at least one of the three desiderata. In particular, S2PC techniques guarantee end-to-end privacy but have either low recall or quadratic cost. In contrast, no end-to-end privacy guarantee has been formalized for solutions that achieve subquadratic cost. This is true even for solutions that compose DP and S2PC: DP does not permit the release of any exact information about the databases, while S2PC algorithms for PRL allow the release of matching records. In light of this deficiency, we propose a novel privacy model, called output constrained differential privacy, that shares the strong privacy protection of DP, but allows for the truthful release of the output of a certain function applied to the data. We apply this to PRL, and show that protocols satisfying this privacy model permit the disclosure of the true matching records, but their execution is insensitive to the presence or absence of a single non-matching record. We find that prior work that combine DP and S2PC techniques even fail to satisfy this end-to-end privacy model. Hence, we develop novel protocols that provably achieve this end-to-end privacy guarantee, together with the other two desiderata of PRL.

cs.DB↗

arXiv2016

Online Event Integration with StoryPivot

Modern data integration systems need to process large amounts of data from a variety of data sources and with real-time integration constraints. They are not only employed in enterprises for managing internal data but are also used for a variety of web services that use techniques such as entity resolution or data cleaning in live systems. In this work, we discuss a new generation of data integration systems that operate on (un-)structured data in an online setting, i.e., systems which process continuously modified datasets upon which the integration task is based. We use as an example of such a system an online event integration system called StoryPivot. It observes events extracted from news articles in data sources such as the 'Guardian' or the 'Washington Post' which are integrated to show users the evolution of real-world stories over time. The design decisions for StoryPivot are influenced by the trade-off between maintaining high quality integration results while at the same time building a system that processes and integrates events in near real-time. We evaluate our design decisions with experiments on two real-world datasets and generalize our findings to other data integration tasks that have a similar system setup.

cs.DB↗

arXiv2016

Effective and Complete Discovery of Order Dependencies via Set-based Axiomatization

Integrity constraints (ICs) provide a valuable tool for expressing and enforcing application semantics. However, formulating constraints manually requires domain expertise, is prone to human errors, and may be excessively time consuming, especially on large datasets. Hence, proposals for automatic discovery have been made for some classes of ICs, such as functional dependencies (FDs), and recently, order dependencies (ODs). ODs properly subsume FDs, as they can additionally express business rules involving order; e.g., an employee never has a higher salary while paying lower taxes compared with another employee. We address the limitations of prior work on OD discovery which has factorial complexity in the number of attributes, is incomplete (i.e., it does not discover valid ODs that cannot be inferred from the ones found) and is not concise (i.e., it can result in "redundant" discovery and overly large discovery sets). We improve significantly on complexity, offer completeness, and define a compact canonical form. This is based on a novel polynomial mapping to a canonical form for ODs, and a sound and complete set of axioms (inference rules) for canonical ODs. This allows us to develop an efficient set-containment, lattice-driven OD discovery algorithm that uses the inference rules to prune the search space. Our algorithm has exponential worst-case time complexity in the number of attributes and linear complexity in the number of tuples. We prove that it produces a complete, minimal set of ODs (i.e., minimal with regards to the canonical representation). Finally, using real and synthetic datasets, we experimentally show orders-of-magnitude performance improvements over the current state-of-the-art algorithm and demonstrate effectiveness of our techniques.

cs.DB↗

arXiv2015

Robust Group Linkage

We study the problem of group linkage: linking records that refer to entities in the same group. Applications for group linkage include finding businesses in the same chain, finding conference attendees from the same affiliation, finding players from the same team, etc. Group linkage faces challenges not present for traditional record linkage. First, although different members in the same group can share some similar global values of an attribute, they represent different entities so can also have distinct local values for the same or different attributes, requiring a high tolerance for value diversity. Second, groups can be huge (with tens of thousands of records), requiring high scalability even after using good blocking strategies. We present a two-stage algorithm: the first stage identifies cores containing records that are very likely to belong to the same group, while being robust to possible erroneous values; the second stage collects strong evidence from the cores and leverages it for merging more records into the same group, while being tolerant to differences in local values of an attribute. Experimental results show the high effectiveness and efficiency of our algorithm on various real-world data sets.

cs.DB↗