Interval Neural Networks: Uncertainty Scores
This work addresses the need for reliable uncertainty estimation in sensitive applications like healthcare, offering an incremental improvement over existing methods with added directional information.
The paper tackles the problem of uncertainty quantification in deep neural networks by proposing interval neural networks (INNs), a fast, non-Bayesian method that produces interpretable uncertainty scores with theoretical justification, demonstrating practical utility in image reconstruction tasks compared to state-of-the-art methods.
We propose a fast, non-Bayesian method for producing uncertainty scores in the output of pre-trained deep neural networks (DNNs) using a data-driven interval propagating network. This interval neural network (INN) has interval valued parameters and propagates its input using interval arithmetic. The INN produces sensible lower and upper bounds encompassing the ground truth. We provide theoretical justification for the validity of these bounds. Furthermore, its asymmetric uncertainty scores offer additional, directional information beyond what Gaussian-based, symmetric variance estimation can provide. We find that noise in the data is adequately captured by the intervals produced with our method. In numerical experiments on an image reconstruction task, we demonstrate the practical utility of INNs as a proxy for the prediction error in comparison to two state-of-the-art uncertainty quantification methods. In summary, INNs produce fast, theoretically justified uncertainty scores for DNNs that are easy to interpret, come with added information and pose as improved error proxies - features that may prove useful in advancing the usability of DNNs especially in sensitive applications such as health care.