Back to search

FRINATEK-Fri prosj.st. mat.,naturv.,tek

Beyond Worst-Case Analysis in Algorithms

Alternative title: Beyond Worst-Case Analysis in Algorithms

Awarded: NOK 12.0 mill.

Modern theoretical computer science faces a fundamental challenge: the current methods fail to explain the effectiveness of modern machine learning algorithms. To reconcile the algorithmic intractability with machine learning, we will develop novel algorithmic and complexity methods. The successful completion of our program will yield progress in both areas of theoretical computer science and machine learning.

The field of theoretical computer science faces a fundamental challenge: the worst-case analysis, the established framework to estimate the computational complexity of problems, fails to explain the effectiveness of modern machine learning algorithms. To address this fundamental challenge, we will revise the foundations of computer science by moving beyond worst-case analysis. We will develop novel algorithmic and complexity methods and use these methods to reconcile the worst-case algorithmic intractability with machine learning. The successful completion of our program will yield progress in both areas of theoretical computer science and machine learning, and hence, in almost every area of science and technology.

Activity:

FRINATEK-Fri prosj.st. mat.,naturv.,tek