SECTION 3.3. A BENEFIT, SINCE IT GIVES MORE OPPORTUNITY FOR ENFORC-

4.1 Evaluation 1

We created a fixed question set of 50 questions of

For the evaluation, which consisted of determining if

the form “What is the capital of X?”, for each state

the resulting top answer was right or wrong, it meant

in the U.S. The inverted question “What state is Z

ultimately deciding on one of three possible out-

the capital of?” was correctly generated in each

comes: the original top answer, the original second

case. We evaluated against two corpora: the

answer, or nil. We hoped to promote a significant

AQUAINT corpus, of a little over a million news-

number of second-place finishers to top place and

wire documents, and the CNS corpus, with about

introduce some nils, with minimal disturbance of