D12-1129 |
performance is typically lower than
|
supervised systems
|
. On the other hand , supervised
|
D12-1099 |
plan to compare our system with
|
supervised systems
|
to identify the gap between the
|
D11-1089 |
splitting noun compounds . The three
|
supervised systems
|
performed almost equally for
|
D11-1085 |
supervised and semi-supervised . The
|
supervised system
|
( " Sup " ) carries out discriminative
|
D11-1090 |
have been developed , and pure
|
supervised systems
|
can provide strong performance
|
D10-1029 |
baselines and substantially below the
|
supervised system
|
. We confirmed our Morfessor
|
D08-1068 |
outperforms a state-of-the art
|
supervised system
|
( McCallum & Wellner , 2005
|
D08-1068 |
compare our system to two recent
|
supervised systems
|
, Ng ( 2005 ) and Denis &
|
D08-1068 |
system that is as accurate as
|
supervised systems
|
. It performs joint inference
|
D09-1071 |
can be used instead of that of a
|
supervised system
|
. 5 Experiments In all our experiments
|
D08-1062 |
and recall of state-of-the-art
|
supervised systems
|
. However , we show better performance
|
D12-1129 |
systems . On the other hand ,
|
supervised systems
|
fall short of carrying out high-performance
|
C02-1112 |
Yarowsky , 2001 ) . Most of the
|
supervised systems
|
used only a basic set of local
|
D09-1002 |
Considering how the performance of
|
supervised systems
|
degrades on out-of-domain data
|
D10-1056 |
It assigns a lower score to the
|
supervised system
|
than to brown , indicating that
|
D10-1056 |
a new language or domain than
|
supervised systems
|
. To address this question ,
|
D09-1120 |
comparable to state-of-the-art
|
supervised systems
|
. <title> Descriptive and Empirical
|
D09-1071 |
human annotation or by a good
|
supervised system
|
. Similar extrinsic evaluation
|
D08-1068 |
suited for joint inference in
|
supervised systems
|
( e.g. , transitivity , which
|
D09-1120 |
) . Both of these systems were
|
supervised systems
|
discriminatively trained to maximize
|