Regression Prior Networks

Sep 28, 2020 (edited Mar 05, 2021)ICLR 2021 Conference Blind SubmissionReaders: Everyone
• Reviewed Version (pdf): https://openreview.net/references/pdf?id=zttNivMlGv
• Keywords: uncertainty, prior networks, regression, ensemble distribution distillation, depth estimation.
• Abstract: Prior Networks are a recently developed class of models which yield interpretable measures of uncertainty and have been shown to outperform state-of-the-art ensemble approaches on a range of tasks. They can also be used to distill an ensemble of models via \emph{Ensemble Distribution Distillation} (EnD$^2$), such that its accuracy, calibration and uncertainty estimates are retained within a single model. However, Prior Networks have so far been developed only for classification tasks. This work extends Prior Networks and EnD$^2$ to regression tasks by considering the Normal-Wishart distribution. The properties of Regression Prior Networks are demonstrated on synthetic data, selected UCI datasets and a monocular depth estimation task, where they yield performance competitive with ensemble approaches.
• One-sentence Summary: Development of Prior Networks and Ensemble Distribution Distillation for Regression Tasks
• Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics
22 Replies