Incorporating Interpretable Output Constraints in Bayesian Neural Networks
This addresses the need for more interpretable and constraint-aware models in high-stakes domains, representing an incremental improvement over typical BNNs.
The paper tackles the problem of incorporating task-specific constraints like safety and fairness into Bayesian neural networks (BNNs), introducing a probabilistic framework that enables interpretable output constraints and demonstrates efficacy on real-world datasets in healthcare, criminal justice, and credit scoring.
Domains where supervised models are deployed often come with task-specific constraints, such as prior expert knowledge on the ground-truth function, or desiderata like safety and fairness. We introduce a novel probabilistic framework for reasoning with such constraints and formulate a prior that enables us to effectively incorporate them into Bayesian neural networks (BNNs), including a variant that can be amortized over tasks. The resulting Output-Constrained BNN (OC-BNN) is fully consistent with the Bayesian framework for uncertainty quantification and is amenable to black-box inference. Unlike typical BNN inference in uninterpretable parameter space, OC-BNNs widen the range of functional knowledge that can be incorporated, especially for model users without expertise in machine learning. We demonstrate the efficacy of OC-BNNs on real-world datasets, spanning multiple domains such as healthcare, criminal justice, and credit scoring.