tech,11-1-J05-1003,ak which rerank the output of an existing <term> probabilistic parser </term> . The <term> base parser </term> produces
tech,1-2-J05-1003,ak <term> probabilistic parser </term> . The <term> base parser </term> produces a set of <term> candidate
other,7-2-J05-1003,ak base parser </term> produces a set of <term> candidate parses </term> for each <term> input sentence </term>
other,11-2-J05-1003,ak <term> candidate parses </term> for each <term> input sentence </term> , with associated <term> probabilities
other,16-2-J05-1003,ak input sentence </term> , with associated <term> probabilities </term> that define an initial <term> ranking
other,21-2-J05-1003,ak probabilities </term> that define an initial <term> ranking </term> of these <term> parses </term> . A second
other,24-2-J05-1003,ak initial <term> ranking </term> of these <term> parses </term> . A second <term> model </term> then
model,2-3-J05-1003,ak these <term> parses </term> . A second <term> model </term> then attempts to improve upon this
other,10-3-J05-1003,ak attempts to improve upon this initial <term> ranking </term> , using additional <term> features </term>
other,14-3-J05-1003,ak <term> ranking </term> , using additional <term> features </term> of the <term> tree </term> as evidence
other,17-3-J05-1003,ak additional <term> features </term> of the <term> tree </term> as evidence . The strength of our
other,10-4-J05-1003,ak of our approach is that it allows a <term> tree </term> to be represented as an arbitrary
other,19-4-J05-1003,ak represented as an arbitrary set of <term> features </term> , without concerns about how these
other,26-4-J05-1003,ak , without concerns about how these <term> features </term> interact or overlap and without the
other,37-4-J05-1003,ak overlap and without the need to define a <term> derivation </term> or a <term> generative model </term>
model,40-4-J05-1003,ak define a <term> derivation </term> or a <term> generative model </term> which takes these <term> features </term>
other,45-4-J05-1003,ak generative model </term> which takes these <term> features </term> into account . We introduce a new
other,7-5-J05-1003,ak We introduce a new method for the <term> reranking task </term> , based on the <term> boosting approach
tech,13-5-J05-1003,ak reranking task </term> , based on the <term> boosting approach to ranking problems </term> described in Freund et al. ( 1998
tech,3-6-J05-1003,ak Freund et al. ( 1998 ) . We apply the <term> boosting method </term> to parsing the <term> Wall Street Journal
skrij podrobnosti