#10050Despite much recent progress on accurate semantic role labeling, previous work has largely used independent classifiers , possibly combined with separate label sequence models via Viterbi decoding.
other,21-2-P05-1073,ak
is a joint structure , with strong
<term>
dependencies
</term>
between
<term>
arguments
</term>
. We
#10084This stands in stark contrast to the linguistic observation that a core argument frame is a joint structure, with strong dependencies between arguments.
other,23-2-P05-1073,ak
strong
<term>
dependencies
</term>
between
<term>
arguments
</term>
. We show how to build a joint
<term>
#10086This stands in stark contrast to the linguistic observation that a core argument frame is a joint structure, with strong dependencies between arguments .
other,14-3-P05-1073,ak
frames
</term>
, incorporating novel
<term>
features
</term>
that model these interactions into
#10102We show how to build a joint model of argument frames, incorporating novel features that model these interactions into discriminative log-linear models.
other,11-4-P05-1073,ak
error reduction
</term>
of 22 % on all
<term>
arguments
</term>
and 32 % on
<term>
core arguments
</term>
#10123This system achieves an error reduction of 22% on all arguments and 32% on core arguments over a state-of-the art independent classifier for gold-standard parse trees on PropBank.