Causality in Databases

Description

When analyzing data sets, users are often interested in the causes of their observations: "What caused my personalized newsfeed to contain more than 10 items related to volcanos?", "Why can't I find any flights with my search criteria?". Database research that addresses these or similar questions is mainly work on lineage of query results, such as why or where provenance, and very recently, explanations for non-answers. While these approaches differ over what the response to questions should be, all of them seem to be linked through a common underlying theme: understanding causal relationships in databases.

Causal relationships cannot be explicitly modeled in current database systems, which offer no specific support for such queries. Mining techniques can infer statistically significant data patterns but they are not sufficient to draw conclusions, as correlation does not necessarily imply causation. The goal of this project is to extend the capabilities of current database systems by incorporating to them causal reasoning. This will allow databases to model causal dependencies, and users to issue queries that can interpret them to provide explanations for their observations. Starting from the very basic functionality of justifying the presence or absence of results for a given query, causality-enabled databases can find many practical applications. For an intuitive introduction and several motivating examples see the Data Bulletin article.

A former member of the project, Alexandra Meliou, has a nice Website on this project http://people.cs.umass.edu/~ameli/projects/causality/

Supported by:

Members

Web Page:

Publications