Obsidian/5 Thesis/5 Resources/Literature_Reviews/A Review of Formal Methods applied to Machine Learning.md

The quick brown fox jumps over the lazy dog. The dog stays blissfully asleep. :)

Robert recommended this topic space to me: formal methods applied to machine learning algorithims. The purpose is clear-cut: if we're able to use formal methods to prove things about machine learning based systems, we can make claims that their implementation in critical systems is safe and viable. This review seems to talk a good bit about that.

First Pass

• Machine learned software is especially viable because it is more general than other types of programming. Things such as computer vision and decision making are very difficult to program manually relative to their ML counterpart.