The YodaQA open source question answering system is a work in progress, but it already represents an easy to deploy end-to-end pipeline for both practical use and research in the field of question answering.
YodaQA has been a topic of several papers and reports at this point, linked in the right column. If you want to know how does YodaQA work, go explore them! If you want to explore the code, besides the top links on the right, Design Notes and UIMA Intro docs are useful to read first.
If YodaQA helped you in your research, please cite:
BAUDIŠ, Petr. YodaQA: A Modular Question Answering System Pipeline. In: POSTER 2015 — 19th International Student Conference on Electrical Engineering. Prague: CTU, 2015.
In science, the key question is to quantitatively answer: How well does it work? We constantly benchmark YodaQA on various datasets. Right now, we focus on the movie dataset – help us build it!. All our datasets live in our GitHub project (look for repos called dataset-factoid-...). Current evaluation notes.