@inproceedings{Carvalho:2020,
author = {Carvalho, EDC and Clark, R and Nicastro, A and Kelly, PHJ},
pages = {12003--12013},
publisher = {IEEE},
title = {Scalable uncertainty for computer vision with functional variationalinference},
url = {http://arxiv.org/abs/2003.03396v1},
year = {2020}
}

Download

TY  - CPAPER
AB  - As Deep Learning continues to yield successful applications in ComputerVision, the ability to quantify all forms of uncertainty is a paramountrequirement for its safe and reliable deployment in the real-world. In thiswork, we leverage the formulation of variational inference in function space,where we associate Gaussian Processes (GPs) to both Bayesian CNN priors andvariational family. Since GPs are fully determined by their mean and covariancefunctions, we are able to obtain predictive uncertainty estimates at the costof a single forward pass through any chosen CNN architecture and for anysupervised learning task. By leveraging the structure of the induced covariancematrices, we propose numerically efficient algorithms which enable fasttraining in the context of high-dimensional tasks such as depth estimation andsemantic segmentation. Additionally, we provide sufficient conditions forconstructing regression loss functions whose probabilistic counterparts arecompatible with aleatoric uncertainty quantification.
AU  - Carvalho,EDC
AU  - Clark,R
AU  - Nicastro,A
AU  - Kelly,PHJ
EP  - 12013
PB  - IEEE
PY  - 2020///
SP  - 12003
TI  - Scalable uncertainty for computer vision with functional variationalinference
UR  - http://arxiv.org/abs/2003.03396v1
UR  - https://openaccess.thecvf.com/content_CVPR_2020/html/Carvalho_Scalable_Uncertainty_for_Computer_Vision_With_Functional_Variational_Inference_CVPR_2020_paper.html
UR  - http://hdl.handle.net/10044/1/79364
ER  - 

Download