arXiv subjects

Daniel Deutch

Publications and source records attributed to Daniel Deutch.

24 records · Page 2Linked to original sources

arXiv2018

Explaining Queries over Web Tables to Non-Experts

Designing a reliable natural language (NL) interface for querying tables has been a longtime goal of researchers in both the data management and natural language processing (NLP) communities. Such an interface receives as input an NL question, translates it into a formal query, executes the query and returns the results. Errors in the translation process are not uncommon, and users typically struggle to understand whether their query has been mapped correctly. We address this problem by explaining the obtained formal queries to non-expert users. Two methods for query explanations are presented: the first translates queries into NL, while the second method provides a graphic representation of the query cell-based provenance (in its execution on a given table). Our solution augments a state-of-the-art NL interface over web tables, enhancing it in both its training and deployment phase. Experiments, including a user study conducted on Amazon Mechanical Turk, show our solution to improve both the correctness and reliability of an NL interface.

cs.CL↗

arXiv2018

Possible and Certain Answers for Queries over Order-Incomplete Data

To combine and query ordered data from multiple sources, one needs to handle uncertainty about the possible orderings. Examples of such "order-incomplete" data include integrated event sequences such as log entries, lists of properties (e.g., hotels and restaurants) ranked by an unknown function reflecting relevance or customer ratings, and documents edited concurrently with an uncertain order on edits. This paper introduces a query language for order-incomplete data, based on the positive relational algebra with order-aware accumulation. We use partial orders to represent order-incomplete data, and study possible and certain answers for queries in this context. We show that these problems are respectively NP-complete and coNP-complete, but identify many tractable cases depending on the query operators or input partial orders.

cs.DB↗

arXiv2016

Query By Provenance

To assist non-specialists in formulating database queries, multiple frameworks that automatically infer queries from a set of examples have been proposed. While highly useful, a shortcoming of the approach is that if users can only provide a small set of examples, many inherently different queries may qualify, and only some of these actually match the user intentions. Our main observation is that if users further explain their examples, the set of qualifying queries may be significantly more focused. We develop a novel framework where users explain example tuples by choosing input tuples that are intuitively the "cause" for their examples. Their explanations are automatically "compiled" into a formal model for explanations, based on previously developed models of data provenance. Then, our novel algorithms infer conjunctive queries from the examples and their explanations. We prove the computational efficiency of the algorithms and favorable properties of inferred queries. We have further implemented our solution in a system prototype with an interface that assists users in formulating explanations in an intuitive way. Our experimental results, including a user study as well as experiments using the TPC-H benchmark, indicate the effectiveness of our solution.

cs.DB↗

arXiv2011

Putting Lipstick on Pig: Enabling Database-style Workflow Provenance

Workflow provenance typically assumes that each module is a "black-box", so that each output depends on all inputs (coarse-grained dependencies). Furthermore, it does not model the internal state of a module, which can change between repeated executions. In practice, however, an output may depend on only a small subset of the inputs (fine-grained dependencies) as well as on the internal state of the module. We present a novel provenance framework that marries database-style and workflow-style provenance, by using Pig Latin to expose the functionality of modules, thus capturing internal state and fine-grained dependencies. A critical ingredient in our solution is the use of a novel form of provenance graph that models module invocations and yields a compact representation of fine-grained workflow provenance. It also enables a number of novel graph transformation operations, allowing to choose the desired level of granularity in provenance querying (ZoomIn and ZoomOut), and supporting "what-if" workflow analytic queries. We implemented our approach in the Lipstick system and developed a benchmark in support of a systematic performance evaluation. Our results demonstrate the feasibility of tracking and querying fine-grained workflow provenance.

cs.DB↗

arXiv2011

On the Limitations of Provenance for Queries With Difference

The annotation of the results of database transformations was shown to be very effective for various applications. Until recently, most works in this context focused on positive query languages. The provenance semirings is a particular approach that was proven effective for these languages, and it was shown that when propagating provenance with semirings, the expected equivalence axioms of the corresponding query languages are satisfied. There have been several attempts to extend the framework to account for relational algebra queries with difference. We show here that these suggestions fail to satisfy some expected equivalence axioms (that in particular hold for queries on "standard" set and bag databases). Interestingly, we show that this is not a pitfall of these particular attempts, but rather every such attempt is bound to fail in satisfying these axioms, for some semirings. Finally, we show particular semirings for which an extension for supporting difference is (im)possible.

cs.DB↗

arXiv2011

Provenance for Aggregate Queries

We study in this paper provenance information for queries with aggregation. Provenance information was studied in the context of various query languages that do not allow for aggregation, and recent work has suggested to capture provenance by annotating the different database tuples with elements of a commutative semiring and propagating the annotations through query evaluation. We show that aggregate queries pose novel challenges rendering this approach inapplicable. Consequently, we propose a new approach, where we annotate with provenance information not just tuples but also the individual values within tuples, using provenance to describe the values computation. We realize this approach in a concrete construction, first for "simple" queries where the aggregation operator is the last one applied, and then for arbitrary (positive) relational algebra queries with aggregation; the latter queries are shown to be more challenging in this context. Finally, we use aggregation to encode queries with difference, and study the semantics obtained for such queries on provenance annotated databases.

cs.DB↗