### Table 1. Incremental don apos;t care search for the interface veri cation example literals / sets Omega calls CPU seconds

2000

"... In PAGE 13: ... To improve e ciency we could use an integer linear program- ming algorithm or a simpler version of Omega. Table1 summarizes the number of Omega calls (satis ability checks) and the CPU time required to nd the don apos;t cares used by SIS (and the underlying algorithms of Espresso) to obtain the simpli cation. CPU times are for a DEC 3000 with 32 MB of memory.... ..."

Cited by 1

### Table 2. The comparison of Bayes classifier, HMM and OMEGA.

"... In PAGE 6: ... Figure 3 demonstrates the success of OMEGA in this simulation domain. Table2 compares OMEGA with Bayes classifier and HMM. Both HMM and OMEGA significantly outperform Bayes classification, but it is hard to tell who is better between HMM and OMEGA.... ..."

### Table 1: Robust routing algorithm. 1 Let G initially be graph of entire network 2 Loop every T:

"... In PAGE 1: ... 2. ROUTING ALGORITHM Our routing algorithm is summarized in Table1 . Let N be the graph of the entire network and let the graph G initially equal N.... ..."

### Table 3. The comparison of Bayes classifier, HMM and OMEGA.

"... In PAGE 6: ... Eight people drove a vehicle with many sensors from Pittsburgh to Grove City and back. This experiment is more difficult than the simulation one, not only because the data is from real world, but also does it involve traffic conditions as a part of the input The result in Table3 is easy to understand, but the interesting thing is that OMEGA even outperforms HMM which is a more complicated method. Further discussion on the comparison of OMEGA and HMM goes to [Deng, 98].... ..."

### Table 1: Accuracy of the Epsilon and Omega tests

"... In PAGE 8: ...Number of Percentage of access pairs for which the Epsilon test access pairs Stopped Stopped Stopped after Assumed tested after ZIV after SIV Banerjee apos;s tests dependence cholsky 82 0 43 56 0 s99 176 0 15 84 0 interf 1456 28 57 14 0 poteng 777 40 57 2 0 nl lt 127 36 40 6 16 btrix 518 44 24 31 0 yacobi 376 0 8 91 0 dctdx 63 4 39 15 39 radb4 56 64 35 0 0 vpenta 261 71 28 0 0 Table 4: Distribution access pairs between di erent subtests categories We have compared the results of the Epsilon and Omega tests for every pair of the tested array accesses for several programs. Table1 shows the summary of our results. For 7 procedures out of 10 that we investigated the Epsilon test always gives the correct result.... ..."

### Table 4: The BME Algorithm Algorithm BME 1. let SG = (VSG; ESG) denotes the site structure 2. let F = fI1; I2; ; Ing denotes the discovered frequent itemsets 3. for each itemset I 2 F do

2000

"... In PAGE 17: ...1.3 Algorithm and Complexity Analysis Initially, the set of frequent itemsets F is pruned to remove all the proper subsets (lines 3 to 5 of Table4 ). Next, for each remaining itemset, an undirected graph FG is constructed based on the page references within the itemset (line 7).... ..."

Cited by 32

### Table 9. The algorithm A8: Learn H(P)? using EQ and MQ (propositional version). 1. Let k = 1.

### Table 1 The comparison of the accuracy between Bayes classifier and OMEGA

"... In PAGE 5: ....b. But they are different from the others. As in Figure 2.a, their curves were so much higher than the others that they are off the graph. Table1 is the comparison of the accuracy between OMEGA and Bayes classifier. It is obvious that OMEGA outperforms Bayes classifier by a large margin.... ..."