C94-1032 morphological analyzers use rule-based heuristic searches . They usually use a connectivity
C86-1067 possible paths , the following heuristic search method is adopted . First , a
D12-1046 tuned on development data using a heuristic search . Parameters that achieved the
C69-2901 nondeterministic procedures employing heuristic search and evaluation . The rules employed
C04-1166 described in ( Max , 2003b ) . A heuristic search procedure in the space of content
D09-1007 framing CKY item generation as a heuristic search prob - lem . Once we recognize
C90-3040 credit , Ng and Mooney discuss a heuristic search mechanism that is guided by coherence
D09-1007 we do better ? Why not perform heuristic search directly on this search space
D09-1007 heuristics to the nodes and run a heuristic search protocol , say A * , on this
C00-2151 optimization . Both complete and heuristic search methods can be eml ) loyed to
A88-1021 " business " along any path of heuristic searching . Therefore we prefer to attach
D09-1007 nonterminals . A key benefit of the heuristic search viewpoint is that it is well
D09-1070 affixes . This typically involves heuristic search procedures and calibrating multiple
D12-1051 2010 ) . However , the greedy heuristic search algorithms only explore a fraction
D09-1070 parameters that need to be set , and heuristic search procedures are crucial for their
D09-1007 given span . 3 Item Generation as Heuristic Search Refer again to the example in
D10-1091 technique , which is to guide a heuristic search in an hypergraph representing
D09-1007 number of nodes visited by the heuristic search , the average time to decode
D12-1051 is natural to use some greedy heuristic search algorithms for inference in some
A00-2023 optimal solution . A capped N-best heuristic search algorithm on the other hand has
hide detail