### Table 3: Variations of ulp, clp, and plg for di erent values of probability increases when becomes very small because the contribution of the probe packets to the bu er queue length becomes non negligible. Furthermore, if a probe

1993

"... In PAGE 9: ... We let ulp denote the (unconditional) loss probability for the probe packets. Thus, ulp is de ned by ulp = P(rttn = 0) Table3 presents the measured values of ulp for di erent values of (ignore ulp and plg for the moment). The loss... In PAGE 9: ...ility, i.e. clp = P(rttn+1 = 0jrttn = 0) We denote by plg the packet loss gap.2 Table3 presents the measured values of clp and plg for di erent values of (refer to the beginning of the section). We observe that clp is greater than ulp for all values of .... ..."

Cited by 153

### Table 1. The proportion of invertible matrices modulo N.

1999

"... In PAGE 7: ... It follows that the proportion can be considered as constant for dimensions of interest, since those dimensions are high (higher than 200). Table1 gives numerical results. It shows that with non-negligible probability, the public matrix B is invertible Table 1.... ..."

Cited by 19

### Table 3: Eigenvectors for image of a face. For clarity and ease of comparison, only those spherical harmonics with non-negligible coefficient magnitudes (greater than .10) are noted.

2002

Cited by 46

### Table 2: Summaries of the posterior distributions of N for the spina bi da data for all models with posterior probability greater than 0.01. ^ N is a Bayes estimate, minimizing a relative squared error loss function

"... In PAGE 15: ... 3.2 Example : Spina Bi da The results of a Bayesian graphical model analysis of the spina bi da example of Table 1 are given in Table2 and Figure 4. In the gure, the value of P (N j D; Mk)P (Mk j D) is plotted for any model Mk with non-negligible posterior probability.... ..."

### Table 1. Estimated complexity of the attack for variable digest sizes.

2003

"... In PAGE 9: ... Estimated complexity of the attack for variable digest sizes. To help select the best hash size for our purposes, we quote from the experiments by Coron and Naccache [11] in Table1 . Taking the first line as an example, an inter- pretation of the data is that, among at least 236 hash digests, the probability of finding one hash value dividing another is non-negligible.... ..."

Cited by 7

### Table 1 shows for each application, its total footprint and the average and minimum think time (Z) obtained for the think time traces. The cumulative page fault rate graphs for each applications are shown in gures 2, 3. We tried to choose the minimum think time in a period with non-negligible number of page faults compared to the whole trace, so that our results are signi cant.

"... In PAGE 4: ...67 0.16 Table1 : Footprint and think time (average and minimum) for di erent applications we considered too long. On the other hand, perl, an interpreted scripting language, has an average think time of 5 microseconds, even when the global nodes had enough memory to hold all its pages locally.... ..."

### TABLE IV. Polynomial ts to the logarithmic derivatives, ik(x) = a0 + a1x + a2x2 + a3x3 + a4x4 + a5x5, as functions of x log10( =10?10) 2 [0; 1] (see Fig. 2). In most cases, polynomials of degree lt; 5 provide su cently accurate ts. Only non-negligible logarithmic derivatives are tabulated. k

### Table 2 contains the same for nal state corrections if they exist. For some variables, e.g. the Jaquet-Blondel variables, the nal state is treated completely inclusive. Then, in accordance with the Kinoshita-Lee-Nauenberg theorem [40] there is no LLA correction. Furthermore, there are certain second order corrections from the initial state which may be non-negligible [8]:

### Table 1. Expressions of the three symmetrized second-order statistical measures (\tr quot; denote the trace and \det quot; the determinant of a matrix). sampling frequency on 16 bits by an OROS AU22 board. The recording equipment was set up in the corridor of a university, i.e with a non-negligible background noise. The recordings are single-session. 67 speakers took part to the experiments, mostly students. They each recorded approx- imately 3 minutes of speech.

"... In PAGE 1: ... Let M and N denote the number of spectral vectors used to estimate the covariance matrices and mean vectors, and p the dimension of the spectral vectors. The mathe- matical expression of the three measures that we used in our experiments, are given in Table1 . Note that, once the covariance matrices X and Y are inverted, and that their determinant is evaluated, the computation of the measures requires very few operations.... ..."

### Table 6: Eigenvectors for image of a face with the mean image value subtracted out (i.e. ignoring the ambient term). For clarity and ease of comparison, only those spherical harmonics with non- negligible coefficient magnitudes (greater than .10) are noted here.

2002

Cited by 46