Challenges in interpretability of additive models
This addresses interpretability problems for researchers and practitioners using additive models in AI, but it is incremental as it reviews existing issues.
The paper reviews generalized additive models, highlighting nonidentifiability issues and challenges in interpretability, arguing for caution in claims about their suitability for safety-critical applications.
We review generalized additive models as a type of ``transparent'' model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming ``interpretability'' or ``suitability for safety-critical applications'' of such models.