Can We Write Better Algorithms With Machine Learning?

Quanta magazine describes an 'explosion of interest' in what they're calling algorithms with predictions, arguing that machine learning tools 'have, in a real way, rejuvenated research into basic algorithms.'

Machine learning and traditional algorithms are 'two substantially different ways of computing, and algorithms with predictions is a way to bridge the two,' said Piotr Indyk, a computer scientist at the Massachusetts Institute of Technology. 'It's a way to combine these two quite different threads....' In the past few years, researchers have shown how to incorporate algorithms with predictions into scheduling algorithms, chip design and DNA-sequence searches.

In addition to performance gains, the field also advances an approach to computer science that's growing in popularity: making algorithms more efficient by designing them for typical uses.... By ignoring the worst-case scenarios, researchers can design algorithms tailored to the situations they'll likely encounter. For example, while databases currently treat all data equally, algorithms with predictions could lead to databases that structure their data storage based on their contents and uses....

[M]ost of these new structures only incorporate a single machine learning element. Tim Kraska, a computer scientist at MIT, imagines an entire system built up from several separate pieces, each of which relies on algorithms with predictions and whose interactions are regulated by prediction-enhanced components.

'Taking advantage of that will impact a lot of different areas,' Kraska said.

Read more of this story at Slashdot.