Example sentences of "[art] [noun sg] [modal v] " in BNC.
Next pageNo | Sentence |
---|---|
1 | The algorithm may arrive at the wrong one , and miss the true goal with maximum worth . |
2 | Then the algorithm may opt for a particular child C on the basis of the gradient of f at N , but the gradient is misleading and in fact f ( C ) is much less then f(N) . |
3 | The algorithm may not find the ’ best ’ answer . |
4 | Even if an algorithm embodied knowledge that perhaps the training set should be transformed to polar coordinates , still the algorithm would have to search a 2-dimensional continuum of possible centres . |
5 | Thus , the algorithm would backtrack to one of the apparently cheaper paths and extend that one . |
6 | Even if the estimated cost of a branch costing .1 was .0999999 , the algorithm would backtrack , exploring the search space breadth-first . |
7 | The cost of the path would increase to as a word from this region was incorporated , and the algorithm would backtrack to one of the apparently better extensions where the cost was . |
8 | If the estimates of the costs on paths from the penultimate node to the terminal node are greater than the actual costs on each path , as in Fig. 8.1. , then the algorithm might be misled into taking a non-optimal path . |
9 | If you prefer that the algorithm should not invent new weights , but only select existing weights from the parent strings , then the crossover points marked ’ x ’ may only be at the ends of 8-bit sequences . |
10 | In some cases the algorithm will spread the addresses evenly over the allocated storage area , and in the ideal case will have an equal probability of generating any address within that area ; in other cases the existing key order can be used to improve the efficiency of record storage . |
11 | When we wish to retrieve a synonym , the algorithm will give an ‘ incorrect ’ address . |
12 | When all paths have the same cost associated with them , equal cost will be synonymous with equal depth and the algorithm will perform breadth-first . |
13 | In Figure 8.2 the algorithm will initially take the leftmost path but will then backtrack and eventually take the correct , rightmost path , because the estimate for the rightmost path is less than the actual cost of any other path . |
14 | Whenever the cost of such a path increases , the algorithm will backtrack to a better looking node . |
15 | If h* ( n ) is wildly optimistic ( i.e. the cost estimate is essentially 0 ) then the algorithm will be guided by g(n) the cost computed so far . |
16 | Under these conditions , however good the heuristic estimate , the algorithm will keep abandoning paths that fail to live up to their initial promise in favour of untried paths that are promising a little more than they will deliver ( Pearl 1984 ) . |
17 | Equal cost will start to look like equal depth ( i.e. breadth-first search ) and the algorithm will explore a broad band of hypotheses . |
18 | If there are too many hypotheses with the same or very similar scores , and the cost of the path increases with its length then the algorithm will explore the search space on a broad front , regardless of the accuracy of the estimate . |
19 | Provided the correct path does in fact score markedly better than other paths , the algorithm will return the correct solution without exploring the entire search tree . |
20 | If the quality of bottom-up information was good , the algorithm could quickly home in on the correct sequence of words . |
21 | If the algorithm can detect a short path to a goal , and always choose the N from OPEN which is on this path , then it will be efficient . |
22 | These algorithms are most suited to tasks with small branching ratios and reversible operators , and with simple states so that the algorithm can remember several states at a time . |
23 | It depends on contexts in parse trees , and the algorithm can only calculate parse trees if it has enough AND symbols ; so clustering depends on the set of available AND symbols . |
24 | Whenever the algorithm can not decide unambiguously which line to follow ( for example , at the intersection of two roads , or at a railway junction or a river confluence ) then the operator is asked to resolve the ambiguity by making a choice . |
25 | We will call this backwards pruning because , at the point where more than one path has the same successor , the algorithm can look back the way it has come , and mark or retain only the highest scoring path . |
26 | Generally , the tag will contain information about the item such as … |
27 | The promoters hope the tag will cut queues . |
28 | It did not look very neat , he had to admit , but unless the prospective buyer looked down into the cabinet the bodgery would not be discovered . |
29 | The gates would have been locked , and the porter may be a drunkard but he has his orders . ’ |
30 | It did n't seem likely that the porter would turn out to be co-operative . |