Uncertainty Propagation through Trained Deep Neural Networks Using Factor Graphs
This work addresses predictive uncertainty estimation for safety-critical applications, representing an incremental advance over existing propagation techniques.
The paper tackled the problem of estimating aleatoric uncertainty in deep neural networks for safety-critical applications by proposing a novel uncertainty propagation method using factor graphs, achieving statistically significant performance improvements across three datasets and two architectures.
Predictive uncertainty estimation remains a challenging problem precluding the use of deep neural networks as subsystems within safety-critical applications. Aleatoric uncertainty is a component of predictive uncertainty that cannot be reduced through model improvements. Uncertainty propagation seeks to estimate aleatoric uncertainty by propagating input uncertainties to network predictions. Existing uncertainty propagation techniques use one-way information flows, propagating uncertainties layer-by-layer or across the entire neural network while relying either on sampling or analytical techniques for propagation. Motivated by the complex information flows within deep neural networks (e.g. skip connections), we developed and evaluated a novel approach by posing uncertainty propagation as a non-linear optimization problem using factor graphs. We observed statistically significant improvements in performance over prior work when using factor graphs across most of our experiments that included three datasets and two neural network architectures. Our implementation balances the benefits of sampling and analytical propagation techniques, which we believe, is a key factor in achieving performance improvements.