M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. In TCS, 2000. 16 Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Risk Based Lattice Cutting For Segmental Minimum Bayes-Risk.. - Kumar, Byrne (2002) (Correct)
....whose transitions are given as where : is a labeling that keeps track of both the words and their position in . To compute the Levenshtein distance, the possible one symbol edits (insertion, deletion, substitution) and their costs can be readily represented by a one state weighted transducer [6]. This accounts for the position indices on the output string . Furthermore, we can reduce the size of this transducer by including only transductions that map words on the transitions of to the words in the best path . We can now obtain all possible alignments between and by the following ....
M. Mohri, F. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, vol. 231, no. 1, pp. 17--32, 2000.
Towards a Unified Framework for Sub-lexical and Supra-lexical.. - Mou (2002) (Correct)
....etc. Since most speech recognition components and constraints can be represented by FSTs, such as the lexicon, phonological rules, n gram language models, word phone graphs, and N best lists, many researchers have recently been trying to organize recognizers in the framework of FSTs. Mohri et al. [71, 72] at AT T have been using a state of the art, large scale FST library to handle very large vocabulary speech recognition. An FST based summit recognizer has also been developed at MIT [37] In non FST based recognizers, the recognition components are usually implemented separately using di#erent ....
M. Mohri, F. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, pp. 17--32, January 2000.
Pronunciation modeling by sharing Gaussian densities.. - Saraçlar, Nock.. (2000) (19 citations) (Correct)
....experimentation while allowing a search over a large set of likely word sequences. Acoustic model training and lattice rescoring is carried out using the HTK HMM toolkit developed by Young, Jansen, Odell, Ollasen and Woodland (1995) The AT T Weighted Finite State Transducer tools provided by Mohri, Pereira and Riley (2000) are used to manipulate word and phone lattices. Performance measures. The most common measure of performance for an ASR system is WER. State of the art ASR systems achieve 30 35 WER on the Switchboard corpus. The baseline system described above has comparable performance. The best possible WER ....
Mohri, M., Pereira, F. C. N. & Riley, M. (2000). The design principles of a weighted finite state transducer library. Theoretical Computer Science, 231, 17--32, Available from http://www.research.att.com/sw/tools/ fsm/.
A General Weighted Grammar Library - Allauzen, Mohri, Roark Self-citation (Mohri) (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. Theoretical Computer Science, 231:17--32, January 2000. http://www.research.att.com/sw/tools/fsm.
A General Weighted Grammar Library - Cyril Allauzen Mehryar Self-citation (Mohri) (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. Theoretical Computer Science, 231:17--32, January 2000. http://www.research.att.com/sw/tools/fsm.
Weighted Finite-State Transducers in Speech Recognition - Mohri, Pereira, Riley (2000) (28 citations) Self-citation (Mohri Pereira Riley) (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. Theoretical Computer Science, 231:17--32, January 2000.
Rational Kernels - Corinna Cortes Patrick (2002) (4 citations) Self-citation (Mohri) (Correct)
No context found.
Mehryar Mohri, Fernando C. N. Pereira, and Michael Riley. The Design Principles of a Weighted Finite-State Transducer Library. Theoretical Computer Science, 231:17--32, January 2000. http://www.research.att.com/sw/tools/fsm.
A Weight Pushing Algorithm for Large Vocabulary Speech.. - Mehryar Mohri Michael (2001) Self-citation (Mohri Riley) (Correct)
No context found.
Mehryar Mohri, Fernando C. N. Pereira, and Michael Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, vol. 231, pp. 17--32, January 2000.
Unknown - Ffl-Removal Algorithm For Self-citation (Mohri) (Correct)
.... A ffl is: In the case of the tropical semiring and with a non acyclic automaton A ffl , the total complexity of the algorithm is: 4 Remarks and experiments We have fully implemented the ffl removal algorithm presented here and incorporated it in recent versions of the FSM library [8]. For some automata of about a thousand states with large ffl cycles, our implementation is up to 600 times faster than the previous implementation based on a generalization of the FloydWarshall algorithm. An important feature of our algorithm is that it admits a natural on thefly implementation. ....
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. Theoretical Computer Science, 231:17--32, January 2000.
Semiring Frameworks and Algorithms for Shortest-Distance Problems - Mohri (2002) (1 citation) Self-citation (Mohri) (Correct)
.... One can decompose the graph into its strongly connected components, use Floyd Warshall or Gauss Jordan s algorithms for computing the all pairs shortest distances within each strongly connected component and then find all pairs shortest distances by considering the acyclic component graph [30]. But the solution remains impractical in presence of large strongly connected components. One can then have recourse to various approximations of Floyd Warshall and Gauss Jordan algorithms, but such algorithms do not fully exploit the sparsity of the graphs. We have devised an approximate ....
.... series represented by a weighted automaton or a weighted transducer [5, 6, 10, 36, 23] An e#cient implementation of the Generic Single Source Shortest Distance algorithm with various queue disciplines including the generic topological order can be found in the latest version of the FSM library [30]. The complexity of the Generic Topological Single Source Shortest Distance algorithm is not di#erent from that of Generic Single Source ShortestDistance in the general case. But we can show that the complexity of the topological single source shortest distance algorithm is linear when G is ....
M. Mohri, F. C. N. Pereira, M. Riley, The Design Principles of a Weighted Finite-State Transducer Library. Theoretical Computer Science 231 (2000), 17--32.
Weighted Grammar Tools: The GRM Library - Mohri (2000) Self-citation (Mohri) (Correct)
....and efficiency, to support competitive large vocabulary dynamic recognition using automata of more than 1M states and transitions. Generality led us to define the most extended algebraic framework for each algorithm 1 as in a general purpose finite state machine library (FSM library) [Mohri, Pereira and Riley 2000]. The weight set for the rules and grammars used in the GRM library can be an arbitrary semiring. 2 The implementation given for each algorithm is very general: it does not depend on the underlying semiring used. A general implementation also helps avoiding redundancy and reducing the amount of ....
....is a ring that may lack negation. WEIGHTED GRAMMAR TOOLS: THE GRM LIBRARY 21 dynamic activation or non activation of rules, and the use of dynamic lists. Access to these features is essential in spoken dialogue applications. The GRM library is a high level library based on the FSM library [Mohri et al. 2000] of which it is using some of the functions. The format of the automata and transducers used in the GRM library is compatible with that of the FSM library. This allows one to use numerous utilities available with the FSM library to manipulate the machines created by the GRM library. In particular, ....
[Article contains additional citation context not shown here]
Mohri, M., Pereira, F. C. N. and Riley, M.: 2000, The design principles of a weighted finite-state transducer library, Theoretical Computer Science 231, 17--32.
Weighted Pushdown Systems and Weighted Transducers - Lal, Touili, Kidd, Reps (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. In TCS, 2000. 16
Algorithms for Minimum Risk Chunking - Martin Jansche Center (2005) (Correct)
No context found.
Mohri, M., Pereira, F., Riley, M.: The design principles of a weighted finite-state transducer library. Theoretical Computer Science 231 (2000) 17--32
Probabilistic Finite-State Machines - Part II - Vidal, al. (Correct)
No context found.
M. Mohri, F. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, vol. 231, pp. 17--32, 2000.
Probabilistic Finite-State Machines - Part I - Vidal, al. (Correct)
No context found.
M. Mohri, F. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, vol. 231, pp. 17--32, 2000.
Task-Specific Minimum Bayes-Risk Decoding using Learned Edit .. - Izhak Shafran And (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library", Theoretical Computer Science, 231:17-32, 2000, http://www.research.att.com/sw/tools/fsm.
Segmental Minimum Bayes-Risk Decoding for Automatic Speech.. - Goel, Kumar, Byrne (2003) (Correct)
No context found.
M. Mohri, F. Pereira, and M. Riley, "The design principles of a weighted finite-state transducer library," Theoretical Computer Science, vol. 231, no. 1, pp. 17--32, 2000.
Corrective Language Modeling For Large Vocabulary Asr With.. - Perceptron Algorithm Brian (Correct)
No context found.
M. Mohri, F. C. N. Pereira, and M. Riley. The design principles of a weighted finite-state transducer library. Theoretical Computer Science, 231:17--32, January 2000.
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