Results 1 - 10 of 28,971

Table 1 in section 2.1 provides certain aspects that are identified for the distributed domain where the AOP approach should be useful. In addition, the most noticeable examples of crosscutting concerns for distributed applications are the following:

in Emergent Behaviour of Aspects in High Performance and Distributed Computing
by Sasa Subotic, Judith Bishop 2005
"... In PAGE 2: ... Logging, tracing, profiling, policy enforcement, pooling, caching, authentication, authorization and transactional management are some crosscutting concerns that can be nicely addressed with aspects. Table1 identifies certain aspects for certain domains. From the given table we see that aspects could refer to location, communication, synchronization, etc.... In PAGE 5: ...Proceedings of SAICSIT 2005 Table1 in section 2.1 provides certain aspects for an image processing domain.... ..."
Cited by 1

Table 3.1: Syntactic properties With regard to the previous requirements, we only keep the following systems: dn, , s, se, *. We may notice that among these calculi, we get the smallest ones ( s with 7 rule schemes and with 8 rules) and the biggest ones ( * with 24 rules and dn with 25 rules).

in An Efficiency Comparison Between Different Explicit Substitutions Calculi.
by Eric Deplagne

Table 1. Notice the following facts: The individual module are small. Since modules are tightly coupled, the number of states of the wristwatch is much less than the product of the number of states of its components*. For example, the display handler is completely driven by the button interpreter and the number of states of their parallel product is no more than the number of states of the button interpreter itself. The resulting code is small and fast. It only contains actions that are inevitable at run-time. There is no process handling and communication overhead. In practice, a transition takes about 500 microseconds on a SUN 3/60.

in Programming a Digital Watch In ESTEREL v3 Programmation d'une Montre Digitale en ESTEREL v3
by Erard Berry Ecole, En Informatique Et Automatique Sophia-antipolis
"... In PAGE 24: ... These variants are rather trivial and will not be discussed further. Table1 . shows the e ect on the resulting automaton and on the compiling time.... ..."

Table 1. Notice the following facts: The individual module are small. Since modules are tightly coupled, the number of states of the wristwatch is much less than the product of the number of states of its components*. For example, the display handler is completely driven by the button interpreter and the number of states of their parallel product is no more than the number of states of the button interpreter itself. The resulting code is small and fast. It only contains actions that are inevitable at run-time. There is no process handling and communication overhead. In practice, a transition takes about 500 microseconds on a SUN 3/60.

in Programming a Digital Watch In ESTEREL v3
by Gerard Berry, En Informatique Et Automatique Sophia-antipolis
"... In PAGE 24: ... These variants are rather trivial and will not be discussed further. Table1 . shows the e ect on the resulting automaton and on the compiling time.... ..."

Table 3.3: Network size comparison. One notices that TDRNN requires signi - cantly less neurons and connections. performed using ASA while RNN data was taken from previous works[5, 8] using the indirect approach for gradient decent method. The comparison is based on the following criteria:

in Training Synaptic Delays in a Recurrent Neural Network.
by Faculty Of Engineering, Barak Cohen, Barak Cohen, Dr. D. Saad

Table 5.1: Truth-degrees in Using the procedure introduced in gure 3.2 for each possible world separately, the results in Table 5.2 follow: which indicate that (values underlined, table 5.2): 1notice that is not even required to be nite. Obviously it would not be possible to keep track computationally of an in nite set of possible worlds

in Automated Reasoning with Uncertainties
by Flávio S. Correa da Silva

Table 2: A Status feature table mapped from the prime relation. prime relation. One such feature table is presented in the following example. Notice that the feature table presented here is in relevance to only one reference attribute, Status, however, the method can be easily extended to generate feature tables in relevance to more than one reference attribute. Example 4.1 Suppose a frequently inquired data set in the University database collects the information about computing science students and their associated teaching assistant information. Let the prime relation be generalized from a data set obtained by a selection from the relation Student using, Major = \computing science quot;, plus an attribute IsTA to indicate whether the student is a teaching assistant (i.e., whether the tuple is joinable with the Course relation on the teaching assistant attribute). The scheme for the initial data relation is, Student(Sname; Status; Sex; Birth date(Day; Month; Y ear); Birth place(City; Province; Country); GPA; IsTA): Following Algorithm 4.1, the attributes Sname, Day, Month, City, and Province are removed, and the prime relation has the scheme,

in Intelligent Query Answering by Knowledge Discovery Techniques
by Jiawei Han, Yue Huang, Nick Cercone 1996
"... In PAGE 14: ...other attributes at a high level. For example, the generalized rule, all of the teaching assistants are graduate students, can be extracted from Table2 based on the fact that the class grad takes all of the IsTA count. The feature table is especially useful for generation of rules associated with quantitative (statistical) information.... In PAGE 14: ... Example 4.2 Let Table2 be the Status feature table extracted from the prime relation. Suppose the query is to nd a generalized rule which distinguishes graduate and undergraduate CS students born in Canada with excellent GPA (Example 3.... ..."
Cited by 20

Table 2. Error scores of our learner in publicly available regression datasets in addition to the parabola dataset described above. The CPU, Body fat were dowloaded from Cubist URL [6], while Boston housing, and Liver disorders can be found at UCI Repositoty [1]. The number of steps followed by GNG was the default value, i.e., the number of training comparisons divided by 10. Notice that LACE reached only 0.03% errors when N was 3 in the parabola dataset.

in Learning to Assess From Pair-Wise Comparisons
by J. Díez, J.J. del Coz, O. Luaces, F. Goyache, J. Alonso, A. M. Peña, A. Bahamonde

Table 3.1 compares the advantages and disadvantages between ASICs, DSPs, and GP processors with multimedia extension. Please notice that pipelining, superscalar and VLIW could be common architectures for all three implementations and therefore are not included in the comparison. The following comparisons are made based on handheld or portable devices. However, the result could be different in terms of other devices.

in Multimedia Extension to a Reconfigurable Processor
by I-kai Wang 2003

Table 1 lists the L2-gain computed using both the standard and new LPV algorithms. The LMI apos;s (12) from the stan- dard LPV method on the gridding points are not feasible for constant P. With the extra variable W( ) in the LMI apos;s, we obtain signi cantly less conservative numbers. Also notice that allowing P to be parameter-dependent greatly reduces conservatism. When 1 is used, the coe cient matrices Zi returned by the optimization are as follows:

in Nonlinear H1 Control: An Enhanced Quasi-LPV Approach
by Yun Huang And, Yun Huang, Ali Jadbabaie 1999
"... In PAGE 5: ...Method Constant P 1 2 3 LPV w/o W( ) not feasible 73:6 73:2 71:4 LPV w/ W( ) 66:4 8:2 8:0 7:4 Table1 : L2-gain of the closed-loop system of Example 3.4 with the given state feedback law Note that the above basis are Legendre polynomials on the operating regions.... ..."
Cited by 2
Results 1 - 10 of 28,971