N06-2009 to the error rate reduction . Paraphrase Selection Module . To select a para - phrase
W11-2504 adaptation of specific domain of paraphrases selection . 6.3 Limitations of MonoDS Implementation
J14-4003 all three approaches learn the paraphrase selection step from data , they rely on
N06-2009 Experimental Results We trained the paraphrase selection module using our QA system ,
W11-2504 a threshold for high precision paraphrase selection . 1 Introduction Paraphrasing
W05-1605 effect of polarity filtering and of paraphrase selection in that system . 6.1 The effect
W05-1605 of these optimisations supports paraphrase selection . 1 Introduction As is well known
P07-1042 realiser with a symbolic approach to paraphrase selection . We now compare it to existing
P07-1042 the sentence John often runs . 4 Paraphrase selection The surface realisation algorithm
W05-1605 augmented the realiser with a paraphrase selection mechanism based on the polarity
W05-1605 technique . 6.2 Paraphrase selection Paraphrase selection permits reducing the combinatorics
W05-1605 the filtering technique . 6.2 Paraphrase selection Paraphrase selection permits
W11-1610 refining of paraphrase sets improves paraphrase selection and is a reasonable alternative
J14-4003 reranking configurations . Beam = paraphrase selection beam ( % of first best probability
W05-1605 automaton path is ensured . 5 Paraphrase selection As pointed out in the introduction
W10-0805 evaluation of translation . 4.1 Paraphrase Selection for Translation Based on the
W05-1605 realised by a clefted structure . Paraphrase selection is implemented by parameterising
N06-2009 feature extraction module , and a paraphrase selection module are exe - cuted . The
N06-2009 2000 ) . We initially formulated paraphrase selection as a three-way classification
N06-2009 present our initial approach to paraphrase selection which shows that , despite the
hide detail