SUPN2 (8).pdf (8.54 MB)
Learning structured gaussians to approximate deep ensembles
conference contribution
posted on 2023-06-10, 04:19 authored by Ivor SimpsonIvor Simpson, Sara Vicente, Neil D F CampbellThis paper proposes using a sparse-structured multivariate Gaussian to provide a closed-form approximator for the output of probabilistic ensemble models used for dense image prediction tasks. This is achieved through a convolutional neural network that predicts the mean and covariance of the distribution, where the inverse covariance is parameterised by a sparsely structured Cholesky matrix. Similarly to distillation approaches, our single network is trained to maximise the probability of samples from pretrained probabilistic models, in this work we use a fixed ensemble of networks. Once trained, our compact representation can be used to efficiently draw spatially correlated samples from the approximated output distribution. Importantly, this approach captures the uncertainty and structured correlations in the predictions explicitly in a formal distribution, rather than implicitly through sampling alone. This allows direct introspection of the model, enabling visualisation of the learned structure. Moreover, this formulation provides two further benefits: estimation of a sample probability, and the introduction of arbitrary spatial conditioning at test time. We demonstrate the merits of our approach on monocular depth estimation and show that the advantages of our approach are obtained with comparable quantitative performance.
History
Publication status
- Published
File Version
- Accepted version
Journal
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)ISSN
2575-7075Publisher
IEEE/CVFExternal DOI
Page range
1-9Event name
IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR)Event location
New Orleans, Louisiana, USAEvent type
conferenceEvent date
19-24 June 2022ISBN
9781665469463Department affiliated with
- Informatics Publications
Full text available
- Yes
Peer reviewed?
- Yes
Legacy Posted Date
2022-07-28First Open Access (FOA) Date
2022-11-01First Compliant Deposit (FCD) Date
2022-07-26Usage metrics
Categories
No categories selectedKeywords
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC