J10-3010 extrinsic evaluation is done by embedding the expansion systems into a
D15-1054 successful application of word embedding techniques for the task of click
H90-1008 rule causes retraversal of the embedding link , and the Past-rule then
D15-1200 relat - edness , controlling for embedding dimen - sionality . We find that
P15-1107 that hierarchically builds an embedding for a paragraph from embeddings
D15-1205 models build a representation ( or embedding ) for a linguistic structure
D14-1062 relevance for the in-domain task . By embedding our latent domain phrase model
J98-4006 especially those involving central embedding ( recursion ) , was one of the
E89-1033 has been implemented , and an embedding of this in an interactive parsing
D15-1054 propose a set of novel joint word embedding methods by leveraging implicit
P15-2108 closest to it in a particular embedding provides a characterization for
H90-1008 predicates to be at about the time of embedding event ( e.g. , the assertion
W15-3822 : Left-Right surrounding based embedding feature ( LR_SBE ) and MAX surrounding
S14-2011 provides dense , low-dimensional embedding for each fragment which allows
D14-1012 effectively incorporating the word embedding features within the framework
W15-1501 on the popular skip-gram word embedding model . The novelty of our approach
D15-1200 highlight the importance of testing embedding models in real applications .
W15-1504 The method , Instance-context embedding ( ICE ) , leverages neural word
D14-1167 continuous vector space . The embedding process attempts to preserve
N06-4008 makes mistakes ) . Ndaona includes embedding and graphics parameter estimation
hide detail