### Table 4. Numbers of nodes in BMDs and MTBDDs for 16-bit precision elementary func- tions.

2006

"... In PAGE 5: ... On the other hand, in MTBDDs for the polynomial func- tions, the numbers of nodes are 2n+1 1, since for all the generated polynomial functions, Lemma 6 holds. Non-polynomial Functions: Table4 compares the num- bers of nodes in BMDs with that in MTBDDs for 16-bit precision elementary functions. As shown in Table 2, for the elementary functions, BMDs require many fewer termi- nal nodes than MTBDDs.... ..."

Cited by 1

### Table 2. Numbers of distinct arithmetic coef- cients for 16-bit precision elementary func- tions.

2006

"... In PAGE 4: ...ounds. This fact veri es the theoretical result (Lemma 2). Table 1 shows that for polynomial functions, many arith- metic coef cients are 0, and many non-zero coef cients have identical values as well. Non-polynomial Functions: In addition to the polyno- mial functions, we represented the non-polynomial elemen- tary functions shown in Table2 . Table 2 compares the num- bers of distinct function values and distinct arithmetic coef- cients for 16-bit precision elementary functions.... In PAGE 4: ... For each elementary function, its domain is 0 x lt; 1 and the func- tion values are rounded to 16-bit precision. Table2 shows that the elementary functions are transformed into the com- pact arithmetic spectrum. For 1 px+1 0:707 and px+1 1, the numbers of distinct function values are smaller than other functions, since their range is smaller than the others.... In PAGE 5: ... Non-polynomial Functions: Table 4 compares the num- bers of nodes in BMDs with that in MTBDDs for 16-bit precision elementary functions. As shown in Table2 , for the elementary functions, BMDs require many fewer termi- nal nodes than MTBDDs. Thus, the total numbers of nodes in BMDs are smaller than that of MTBDDs.... ..."

Cited by 1

### Table 1. RMSE and disparity of estimated inverse response func- tions in terms of normalized irradiance.

"... In PAGE 6: ... In computing disparity, four sequences for each cam- era are used. The result is summarized in Table1 .For the Sony DSR-PD190P and PointGrey DragonFly cameras, the comparison curve is generated by averaging the curves obtained by the method in [16] and by fitting to Macbeth color checker measurements, which were interpolated us- ing a fifth-degree polynomial.... ..."

### Table 7.1 Error bounds for the polynomials of Example 1. Figure 7.1 and the error bounds in Table 7.1 demonstrate that for these examples the func- tions (14) are very close approximations of the lower convex envelopes. Example 2: We consider three univariate polynomials from the literature. The rst two polynomials

### Table 1 displays the preliminary results for both cases in which the number of regions in the optimal configuration was fixed at 11 and when the number was unrestricted. In the 2nd column of the table, the intraregional transplants increase per year is presented for the former case. The intraregional transplants in- crease per year, and the number of regions in the opti- mal regional configuration is presented in the 3rd and 4th columns of the table, respectively, for the latter case. It appears that regardless of the liver viability func- tion chosen (linear or polynomial), the optimal re-

### Table 1: Overall results on the development set. Func- tions with prefix g are gold functions, providing bounds of our performance. The top row is the upper bound per- formance of our architecture. The bottom row is the real performance.

2004

"... In PAGE 4: ... We ran the learning algorithm on the training set (with predicted input syntax) with a polynomial kernel of de- gree 2, for up to 8 epochs. Table1 presents the ob- tained results on the development set, either artificial or real. The second and third rows provide, respectively, the loss suffered because of errors in the filtering and scor- ing layer.... ..."

Cited by 7

### Table 1: Overall results on the development set. Func- tions with prefix g are gold functions, providing bounds of our performance. The top row is the upper bound per- formance of our architecture. The bottom row is the real performance.

2004

"... In PAGE 4: ... We ran the learning algorithm on the training set (with predicted input syntax) with a polynomial kernel of de- gree 2, for up to 8 epochs. Table1 presents the ob- tained results on the development set, either artificial or real. The second and third rows provide, respectively, the loss suffered because of errors in the filtering and scor- ing layer.... ..."

Cited by 7

### Table 2: Storage manager interface func- tions.

1994

"... In PAGE 8: ... The repository-dependent part of a gateway comprises an interface to the local storage manager, and must be hand crafted for each repository. This task, however, is straightfor- ward requiring the implementation of eight basic storage management interface func- tions that are called by the server library (see Table2 ). (Note that a simple read-only gate- way need only implement OpenRepository(), CloseRepository(), and GetObject().... ..."

Cited by 6

### Table 1: An overview of the available analysis func- tions.

1997

"... In PAGE 4: ... Besides, as was explained in Section 2, the library can easily be extended with new functions when this is needed. Table1 lists the analysis functions available in the language. The parameters called pattern refer to pattern-matching strings resulting in a list of one or more numbers.... ..."

Cited by 4

### Table 3. Cost of basic anonymization func- tions

2006

"... In PAGE 7: ....0 GHz, with 512 MB main memory. As input source we used a 2 GB tcpdump trace consisting of HTTP traffic. Our results are summarized in Table3 . The metric we use is user time, measured with the time command-line utility.... ..."

Cited by 1