Verifying text summaries of relational data sets

A method comprising:

• obtaining a text document containing at least one claim about data of a relational database;
• extracting claim keywords from the text document;
• identifying query fragments based at least in part on indexing of one or more relational data sets of the relational database;
• determining relevance of the claim keywords to the query fragments;
• identifying candidate queries based at least in part on probabilistic inferencing from the determined relevance of the claim keywords to the query fragments;
• evaluating the candidate queries against the relational database;
• and
• determining consistency of the claim with the data of the relational database based at least in part on results of the evaluation of the candidate queries;
• wherein the method is performed by a processing platform comprising at least one processing device comprising a processor coupled to a memory.

A computer program product comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device of a processing platform causes processing platform:

• to obtain a text document containing at least one claim about data of a relational database;
• to extract claim keywords from the text document;
• to identify query fragments based at least in part on indexing of one or more relational data sets of the relational database;
• to determine relevance of the claim keywords to the query fragments;
• to identify candidate queries based at least in part on probabilistic inferencing from the determined relevance of the claim keywords to the query fragments;
• to evaluate the candidate queries against the relational database;
• and
• to determine consistency of the claim with the data of the relational database based at least in part on results of the evaluation of the candidate queries.