July 15, 2025
The professor is a very engaging lecturer, although the lectures I feel are relatively short. I’m already starting to see and think about how this can applied to performing large calculations, although I’d like to look more into how these calculations can be done faster. Currently we can define an upper bound of O(n3) for matrix multiplication because we have to perform n2 calculations for each row in the matrix. There are some faster algorithms like Strassen’s and Coppersmith-Winograd’s which reduce the overall time complexity to just under n3.
A cool improvement although the drawback of analyzing algorithms solely by their asymptotic upper bound is that we ignore constant factors which (unfortunately) do affect our real execution speed. I’d like to look more into this though!