Tools
Sorted by:
Try your query at:
 |
 |
 |
 |
 |
 |
Results 1 - 10
of
79,786
Table 1 is based on the behavior of components imple- mented in phase one. We obtained these values by esti- mating the number and types of operations based our understanding gained by implementing and testing the functionality of the workload generators explained in Section 3.1.1. For the predicted behavior for the remain- ing components, please refer to [13].
in Abstract
"... In PAGE 6: ...eet the constraints discussed in section 4.1.1. Table 1 describes the predicted behavior for two of the SLICE components (which were defined using the Workload Modeling Language) to illustrate the various types of workload and actions for a CoWorkEr. Planner -1 CoWorkEr Workload performed every second publish command of size 24 bytes Workload performed after receipt of a track event alloc 30 KB; 55 dbase ops; 45 CPU ops; publish assessment of size 132 bytes; de- alloc 30 KB Configuration-Optimization CoWorkEr Workload performed at startup time alloc 1 KB; 25 dbase ops; 1 CPU ops; 10 dbase ops; dealloc 1 KB Workload performed after receipt of an as- sessment event alloc 5 KB; 40 dbase ops; 1 CPU op; pub- lish command of size 128 bytes; dealloc 5 KB Workload performed after receipt of a status event 1 dbase op Table1... ..."
Table 2: Numbers of patterns found within the soprano line of 185 Bach chorales, using the melodic interval viewpoint. The second column refers to the number of raw, unfiltered patterns occurring in at least k pieces. The third column refers to the the number of statistically significant patterns at a p-value of 0.01. The last column refers to the number of patterns remain- ing at the leafs of the subsumption taxonomy. The average length of the longest patterns is indicated in brackets.
2001
"... In PAGE 5: ... General behavior of the algorithm. Table2 illustrates the behavior of the algorithm as a function of the parameter k (the minimum number of pieces a pattern must occur in). At lower values, the method discovers many patterns.... ..."
Cited by 24
Table 2: Numbers of patterns found within the soprano line of 185 Bach chorales, using the melodic interval viewpoint. The second column refers to the number of raw, unfiltered patterns occurring in at least k pieces. The third column refers to the the number of statistically significant patterns at a p-value of 0.01. The last column refers to the number of patterns remain- ing at the leafs of the subsumption taxonomy. The average length of the longest patterns is indicated in brackets.
"... In PAGE 5: ... General behavior of the algorithm. Table2 illustrates the behavior of the algorithm as a function of the parameter k (the minimum number of pieces a pattern must occur in). At lower values, the method discovers many patterns.... ..."
Table 9: Allowed combinations of fermions and their contribution to the remain- ing anomalies.
Table 4: Overall Equivalence Identification Circuit Equiv. Proven Proven Remaining Remaining
2001
"... In PAGE 5: ... All the equivalent pairs of c499, c880, c1355 and c5315 are already identified by the evaluation technique (with static and dynamic analysis) and the implication technique [11]. The number of equivalent fault pairs identified by using all the techniques proposed here are reported in Table4 and compared with several other results. Column 2 shows the number of equiv- alent pairs obtained using the diagnostic test generation procedure DIATEST [1].... ..."
Cited by 7
Table 6: Comparison with Previous Results on Identifying Equivalences Circuit Equiv. Provenpairs Provenpairs Remaining Remaining % increase
1999
"... In PAGE 5: ... Pairs with intermediate cones of up to 10 inputs were considered in this evaluation. Comparison with Previous Results A comparison with the previously proposed method [2] is given in Table6 . Column 2 shows the number of equiva- lent pairs obtained exhaustively using DIATEST [1].... ..."
Cited by 6
Table 1: Word error rates and average WER-to-baseline ratios for different systems, all based on MFCC features. The first line is the standard HTK GMM baseline defined for the Aurora task. The second line is a conventional hybrid system, based on the posterior estimates generated by the neural-net acoustic model. The remain- ing four lines are the results of tandem systems, feeding versions of the posteriors into the HTK system; logp indicates that the log of the posteriors are taken, whereas lino systems use the neural-net outputs directly, before the final nonlinearity is applied. +KL indicates that a full-rank Karhunen-Loeve orthogonalization was applied before passing the values to the HTK system.
2000
Cited by 45
Table 1: Word error rates and average WER-to-baseline ratios for different systems, all based on MFCC features. The first line is the standard HTK GMM baseline defined for the Aurora task. The second line is a conventional hybrid system, based on the posterior estimates generated by the neural-net acoustic model. The remain- ing four lines are the results of tandem systems, feeding versions of the posteriors into the HTK system; logp indicates that the log of the posteriors are taken, whereas lino systems use the neural-net outputs directly, before the final nonlinearity is applied. +KL indicates that a full-rank Karhunen-Loeve orthogonalization was applied before passing the values to the HTK system.
Table 1: Word error rates and average WER-to-baseline ratios for different systems, all based on MFCC features. The first line is the standard HTK GMM baseline defined for the Aurora task. The second line is a conventional hybrid system, based on the posterior estimates generated by the neural-net acoustic model. The remain- ing four lines are the results of tandem systems, feeding versions of the posteriors into the HTK system; logp indicates that the log of the posteriors are taken, whereas lino systems use the neural-net outputs directly, before the final nonlinearity is applied. +KL indicates that a full-rank Karhunen-Loeve orthogonalization was applied before passing the values to the HTK system.
Results 1 - 10
of
79,786
