
Gunnar Wolf - Nice grey life - page 6
Showing posts 51 – 60





Nowadays, we can assume readers of Computing Reviews are familiar with the ideas behind machine learning, where neural networks are trained with large training sets so that they “learn” to recognize patterns within said dataset; some of those patterns are readily identifiable by a human observer, but some might be deep, very subtle, and impossible for us to understand. Neural networks usually keep learning (and thus adjusting their behaviors) from further data they receive in production. This short article deals with the problem termed as “machine unlearning”: how can we ensure all traces of a given problematic case can be...




subscribe via RSS