Question Answering (QA)
is a widely researched area in Natural Language Processing
. Recent developments even lead to build systems such as IBM Watson
which can compete in Jeopardy
shows. However, the question here is that “did we solve the problem?” and “can we move to a new area?”. The simple answer is “No, and in my POV at least not in next 10 years”. Although we are making a very good progress in QA research, there are hundreds of open issues and hurdles. Few of them are worth mentioning: answer presentation, Question Answering over Linked Data (specifically transforming NL question to SPARQL
), and automatic evaluation in QA (moving beyond MRR).