02-06
0.436
0.157
0.486
11.5%
collection of paired questions and answer sen-
Table 10: Top-1 accuracy of the QuALiM system on
tences. Other than previous work employing de-
its own and when combined with the algorithm de-
pendency paths for QA, our approach does not as-
scribed in this paper, when only considering questions
sume that a valid answer sentence is similar to the
for which a pattern could be acquired from the training
question and it allows many potentially very dif-
data. All increases are statistically significant using asign test (p<0.05).
ferent syntactic answer sentence structures. The
algorithm is evaluated using TREC data, and it
is shown that it outperforms an algorithm that
merely uses the syntactic information contained