J11-3011 class of problems is related to graph algorithms . Graph algorithms can be applied
D10-1012 two simple , yet ef - fective , graph algorithms to induce the senses of our queries
C04-1032 pair . Then , we use efficient graph algorithms to determine the symmetric alignment
D11-1072 candidates are omitted for the graph algorithm . The robustness tests and the
C04-1032 advantage is the efficiency of the graph algorithms used to deter tions ( Verbmobil
E14-4046 summarization . Previous word graph algorithms are based on bigrams . Words
J00-1005 different variants of the per graph algorithm . In our experiments , the per
D10-1073 similarity . The hierarchical random graphs algorithm ( Clauset et al. , 2008 ) was
D15-1110 which enables the use of standard graph algorithms ( like MST ) . Also , this is
D11-1072 solution . 5.3 Robustness Tests The graph algorithm generally performs well . How
H93-1033 directly to justifying these plan graph algorithms in terms of a formal theory of
D12-1035 relations . Moreover , instead of graph algorithms or factor-graph learning , we
E14-1010 apply different semisupervised graph algorithms ( Mincuts , Randomized Mincuts
D12-1128 systems ) . Among the different graph algorithms , PLength consistently outperforms
D11-1072 assigned an entity before running the graph algorithm . In summary , we observed that
D11-1072 various forms of weights . The graph algorithm makes use of Webgraph ( Boldi04
D12-1035 constraints , we do not employ graph algorithms , but model the general disambiguation
D10-1012 V ´ eronis , 2004 ) -- a graph algorithm based on the identification of
D12-1036 similarity score obtained from the graph algorithm over sentences , and degree of
I05-2004 mentioned earlier , different graph algorithms can be used for producing the
hide detail