Abstract
We propose and investigate new complementary methodologies for estimating
predictive variance networks in regression neural networks. We derive a locally
aware mini-batching scheme that result in sparse robust gradients, and show how
to make unbiased weight updates to a variance network. Further, we formulate a
heuristic for robustly fitting both the mean and variance networks post hoc.
Finally, we take inspiration from posterior Gaussian processes and propose a
network architecture with similar extrapolation properties to Gaussian
processes. The proposed methodologies are complementary, and improve upon
baseline methods individually. Experimentally, we investigate the impact on
predictive uncertainty on multiple datasets and tasks ranging from regression,
active learning and generative modeling. Experiments consistently show
significant improvements in predictive uncertainty estimation over
state-of-the-art methods across tasks and datasets.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of 33rd Conference on Neural Information Processing Systems |
| Number of pages | 11 |
| Publication date | 2019 |
| Publication status | Published - 2019 |
| Event | 33rd Conference on Neural Information Processing Systems - Vancouver Convention Centre, Vancouver, Canada Duration: 8 Dec 2019 → 14 Dec 2019 Conference number: 33 https://nips.cc/Conferences/2019/ |
Conference
| Conference | 33rd Conference on Neural Information Processing Systems |
|---|---|
| Number | 33 |
| Location | Vancouver Convention Centre |
| Country/Territory | Canada |
| City | Vancouver |
| Period | 08/12/2019 → 14/12/2019 |
| Internet address |