V.M. Jim'enez, A. Marzal and E. Vidal: "Efficient Enumeration of Sentence Hypotheses in Connected Word Recognition", Proceedings of EUROSPEECH-93, pp. 2183--2186. (1993)  Home/Search   Document Not in Database   Summary   Related Articles

.... dramatically speeds up Shier s algorithm, but the computational complexity remains quadratic in K [SK87, SK89, MC91] A new algorithm for the generation of the K Best Sentence Hypotheses for a given utterance was recently proposed by the authors in the field of Automatic Speech Recognition [JM92, JMV93]. In this paper, we present this algorithm as a general procedure for the K shortest paths problem, and formally prove its correctness and efficiency. 2 Definitions and Problem Formulation Let G = V; E) be a directed graph, where V is the set of nodes and E V Theta V is the set of arcs. Let ....

.... requires visiting 450 nodes in the worst case, whose time cost is negligible when compared to that of the first stage (oneto all shortest path computation) More details about this application, as well as an experimental comparison of time complexity for the different algorithms, can be found in [JM92, JMV93, Marz94]. The Recursive Enumeration Algorithm is also well suited for distributed computation of the K shortest paths in communication networks, since every node can keep and manage its own set of candidates to next shortest path. Finally, the algorithm is easily extensible to find the K shortest paths ....

V.M. Jim'enez, A. Marzal and E. Vidal: "Efficient Enumeration of Sentence Hypotheses in Connected Word Recognition", Proceedings of EUROSPEECH-93, pp. 2183--2186. (1993)

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC