Neural Basis Models for Interpretability

However, these models are typically difficult to train, require numerous parameters, and are difficult to scale.