Efficient tree pattern matching

A classic open problem on tree pattern matching is whether the naive O(mn)-step algorithm for finding all the occurrences of a pattern tree of size m in a text tree of size n can be improved. An O(nM/sup 0.75/ polylog(m))-step algorithm for this tree pattern matching problem is designed. The problems of linear string matching with don't care symbols and linear string max-min convolution are treated.<<ETX>>