Inside Epistatic Nets

Regularizing neural nets for better prediction of biological fitness landscapes

Match your inductive bias to your problem space

All machine learning (ML) algorithms have an inherent inductive bias. Inductive biases are the assumptions we make about the likelihood of encountering certain kinds of solutions. The classic example of an inductive bias is Occam’s Razor, where we assume that simple examples are more likely than complicated ones. In ML, we often translate an …

computational biology deep learning inside machine learning