### Table 1: The dimensions of effort according to Laban as described by Zhao [Zhao].

### Table 2. a taxonomy of human co-presence in a dyadic situation (from Zhao, 2001).

### Table 2 : a taxonomy of human co-presence in a dyadic situation (from Zhao, 2001)

2003

### Table 2. The classical isomerization rate constants from RRKM theory, Gray-Rice theory, Zhao-Rice theory and trajectory calculations (in units of 10?4a.u.).

"... In PAGE 10: ...ial. More speci cly, they should be multiple exponential decay. A simple estimate of the isomerization rate constant can be obtained from the time taken for the initial concentration of the reactant ensemble to decay 95% that is quite close to the one obtained from the best exponential tting. The isomerization rate constants from RRKM theory, Gray-Rice theory, Zhao- Rice theory and trajectory calculations are displayed in Table2 . It shows that for this model system the RRKM theory generates about an order of magnitude overestimate of the reaction rate constant.... ..."

### Table 2: Summary of performance comparison. 12 image chips were used. No OCR was performed for Kamal and Zhao apos;s method due to the poor binarization quality.

1998

"... In PAGE 7: ... 12 image chips cropped from our test image set are used for this experiment. The total number of characters/words are counted as shown in Table2 . Each of these image chips is then binarized using the four algorithms.... In PAGE 7: ... Each output is assigned a score from 0 (worst) to 10 (perfect) based on the cleanness, readability, separability (characters should not be attached to each other) and completeness (a character should not be broken) of the characters. The result is shown in Table2 under column \Quality quot;. To obtain more objective measures of the performances of the algorithms, the binarized images are fed to Caere apos;s WordScan Plus 4.... In PAGE 7: ...0 for OCR. The correctly recognized characters and words are then counted with the totals shown in Table2 under columns \OCRed Char quot; and \OCRed Word quot;. Note that a word is correctly recognized if and only if all its characters are correctly recognized.... In PAGE 7: ... Note that a word is correctly recognized if and only if all its characters are correctly recognized. As shown in Table2 , our algorithm out-performs all the rest of the algorithms with a signi cant margin in both subjective and objective measures. Tsai apos;s method comes second in both categories.... ..."

Cited by 9

### Table 6: Large scale problems arising from the quantum chemistry problem (1):Nakata et al. [8] (2):Zhao et al. [17] size of the Schur complement matrix (= m)) 24,503 7,230

2003

"... In PAGE 9: ...chemistry. The size of the Schur complement matrix of their formulation is not so large, however, the size of matrices X and Y are very large compared with the SDP formu- lation proposed in [8] (See Table6 ). We solved these two problems on the PC cluster which we call ACT-JST cluster.... In PAGE 9: ... Therefore, the current version of the SDPARA is suitable for a large scale SDP problem which involves a large Schur complement matrix and not so large variables like the SDP problem (1) in Table 6. The SDP problem (2) in Table6 requires that all processors need more than 1 GB for input data, variables and the distributed Schur complement matrix. AS a given SDP problem to be solved gets larger, it becomes di cult to share common input data and variables with all processors; we eventually need to distribute input data and variables among the processors to save memory in each processor.... ..."

Cited by 3

### Table 2.1 Computational elements of four hierarchical methods. p denotes the number of terms retained in the GR and Zhao apos;s methods, while K denotes the number of integration points in Anderson apos;s method.

### Table 6: Large scale problems arising from the quantum chemistry problem (1):Nakata et al. [8] (2):Zhao et al. [17] size of the Schur complement matrix (= m)) 24,503 7,230

"... In PAGE 9: ... [17] proposed another SDP formulation of the quantum chemistry. The size of the Schur complement matrix of their formulation is not so large, however, the size of matrices X and Y are very large compared with the SDP formulation proposed in [8] (See Table6 ). We solved these two problems on the PC cluster which we call ACT-JST cluster.... In PAGE 9: ...As we described above, each processor must maintain all input data and variables, whereas the Schur complement matrix is stored on distributed memory. Therefore, the current version of the SDPARA is suitable for a large scale SDP problem which involves a large Schur complement matrix and not so large number of variables like the SDP problem (1) in Table6 . The SDP problem (2) in Table 6 requires that all processors need more than 1 GB for input data, variables and the distributed Schur complement matrix.... ..."