MotherTree: Meta-learning on synthetic data improves decision tree training

Conventional decision tree algorithms produce effective, transparent models that can be audited, communicated, and deployed independently of the training data, but require learning every new task from scratch.