Random commentary about Machine Learning, BigData, Spark, Deep Learning, C++, STL, Boost, Perl, Python, Algorithms, Problem Solving and Web Search
How are we measuring the quality of a solution? If it's the usual worst-case asymptotic time complexity, then you can't go below O(n^2).
how about sorting it?
We can sort it, but this doesn't change the argument that any algorithm will have to read all the matrix elements.Or are you saying that it's somehow sorted initially?