### Table 1. Some affine Eulerian polynomials.

709

"... In PAGE 12: ... Thus the conjecture for tildewide An and tildewide Cn follows from the fact that all roots of the classical Eulerian polynomials are known to be real [13]. Furthermore, using the data in Table1 , it is easy to check that the conjecture holds for the exceptional groups. To collect supporting evidence for the remaining groups tildewide Bn and tildewide Dn, we have determined explicit exponential generating functions for the corresponding affine Eulerian polynomials (see Proposition 6.... ..."

### Table 1. Some affine Eulerian polynomials.

"... In PAGE 12: ... Thus the conjecture for tildewide An and tildewide Cn follows from the fact that all roots of the classical Eulerian polynomials are known to be real [13]. Furthermore, using the data in Table1 , it is easy to check that the conjecture holds for the exceptional groups. To collect supporting evidence for the remaining groups tildewide Bn and tildewide Dn, we have determined explicit exponential generating functions for the corresponding affine Eulerian polynomials (see Proposition 6.... ..."

### TABLE III Comparison of effect of the warp vs. affine on selected variance measures per voxel from the anatomical (MRI) and functional (PET) analysis.

### Table 9: Requirements affinity and opportunity to reuse

"... In PAGE 10: ...Consider the two sets of user requirements for the previously discussed Dice game and a Coin game (see Table9 on the previous page). Assume that dice-game requirements have already been refined and implemented, hence they have all been already classified (see Table 2 and Table 4).... In PAGE 10: ...), we would expect the best matches and identification of reuse opportunities to fall along the matrix diagonal. So there they are (see Table9 ) - clear opportunity to reuse! Requirement C3 does not have a clear match with any of the D1-D7, it matches the majority of artefacts in the repository (dashed line in Table 9). Since both its quot;function quot; and quot;method quot; facets are undefined (see Table 5), they result in small conceptual distances from every other facet value, hence, leading to the high affinity with every design artefact.... In PAGE 10: ...), we would expect the best matches and identification of reuse opportunities to fall along the matrix diagonal. So there they are (see Table 9) - clear opportunity to reuse! Requirement C3 does not have a clear match with any of the D1-D7, it matches the majority of artefacts in the repository (dashed line in Table9 ). Since both its quot;function quot; and quot;method quot; facets are undefined (see Table 5), they result in small conceptual distances from every other facet value, hence, leading to the high affinity with every design artefact.... ..."

### Table 1: Summary of affinity specifications and data distribution constructs.

"... In PAGE 4: ...2 Finally, the home function returns the number of the processor that contains the given object allocated in its local memory. The various affinity hints are summarized in Table1 . If affinity is specified for multiple objects then we currently schedule the task based on the first object.... ..."

### Table 3: Summary of different binding potential measures, their V3 notation, expansion in terms of concentration and affinity of binding sites (the bracketed term on the bottom allows for competition), their calculation and the input function required

in 1

### Table 16. Word error rate comparing systems with different vocabulary sizes. Wall Street Journal was decoded with a small dictionary of 2000 words and a large vocabulary of 20000 words. The small dictionary was derived from the test set, thus the system has no out of vocabulary (OOV) words. The language model weight was set so the baseline WER would be the same for all tasks. The language weight is 0.31 for WSJ with a 2k dictionary, 9.5 for WSJ with a 20k dictionary, and 4.85 for RM1. Normalization results in statistically significantly better performance in all cases. Normalization was performed with an affine function fitting the minimal points and the medians of the three first formants.

"... In PAGE 5: ...zation of the training data becomes irrelevant when less data is available per speaker. . . . 80 Table16... In PAGE 96: ...e., language weights that would actually be used in a real system, Results for the first scenario, with the same baseline word error rates, are presented on Table16 . Results for the second scenario, realistic settings, are presented on Table 17.... ..."

### Table 1 Comparison of the methods to compute affine structures. g, genus of the domain manifold M; b, number of boundaries of M. Method # of singularities Location of singularities Area distortion Angle distortion Transition function

### Table 1 are nearly the same) those procedures that exploit some features associated with the ASIC executor class. Table 1. Affinity average values

2002

"... In PAGE 5: ... During the validation process, the values of the metrics previously defined have been collected, and the affinity value of each functionality has been evaluated in the normalized form. Interesting considerations can be made by analyzing the averages of the affinity values on the whole test suite, for the DSP applications, and for the others (see Table1 ). ADSP for the DSP applications is fairly larger than the other affinity values and the ADSP values evaluated for the other application cases.... ..."

Cited by 6

### Table 1. The number of cosets, weight distribution and autocorrelation spectra of affine equivalent classes of RM(3, 6)/RM(1, 6). The functions are represented in abbreviated notation (only the number of the variables) and the sum should be considered modulo 2.

"... In PAGE 11: ... 5 variables, while the nonlinear part of the functions of type IV depends on 5 or 6 variables. From Table1 in Appendix B, we derive the corresponding ANF. intersectionsq unionsq 3.... ..."