Ask Wikipedia

Title

Authors: Thomas Roche, Aaron Mishkin

What is the problem?

Wikipedia is a massive knowledge base containing information on a wide variety of subjects. Users looking for a single fact using Wikipedia are often directed to pages with significantly more information than they desire. A natural language interface that can understand single sentence questions and query Wikipedia for answers will solve this problem for users. The implementation of this system will rely on the data provided by the WikiData Project, and use Prolog to interpret user input.

What is the something extra?

There are two elements of the project that require investigation for plausibility: parsing arbitrary single-sentence questions, and utilizing the WikiData knowledge base in Prolog. Parsing arbitrary single-sentence questions requires the ability to parse a large number of question formats to extract constraints and other relevant information. It also requires identifying malformed queries and responding robustly. Using the WikiData knowledge necessitates that its information be converted into a format usable by Prolog. It may also be necessary to limit the quantity of information imported from WikiData taking into consideration available computer memory and hard-drive space.

What did we learn from doing this?

(This should be written after you have done the work.) What is the bottom-line? Is logic programming suitable for (part-of) the task? Make sure you include the evidence for your claims.

To be completed after the project has been finished.