Retrospective Uncertainties for Deep Models using Vine Copulas

Nataša Tagasovska, Firat Ozdemir, Axel Brando
2/24/2023
cs.LG

No official code found. Best implementation: tagas/vcnn (3 stars, last commit 2022-12-01).

Abstract

Despite the major progress of deep models as learning machines, uncertainty estimation remains a major challenge. Existing solutions rely on modified loss functions or architectural changes. We propose to compensate for the lack of built-in uncertainty estimates by supplementing any network, retrospectively, with a subsequent vine copula model, in an overall compound we call Vine-Copula Neural Network (VCNN). Through synthetic and real-data experiments, we show that VCNNs could be task (regression/classification) and architecture (recurrent, fully connected) agnostic while providing reliable and better-calibrated uncertainty estimates, comparable to state-of-the-art built-in uncertainty solutions.

View on arXivView PDF

Code Implementations(1)

tagas/vcnnCommunity100%
31Jupyter NotebookDec 1, 20223 years ago

Similar Papers

Uncertainty Quantification for Large-Scale Deep Networks via Post-StoNet Modeling

80% match2025

Uncertainty Quantification of Graph Convolution Neural Network Models of Evolving Processes

80% match2024

Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks

80% match1 repo102023

Variation Due to Regularization Tractably Recovers Bayesian Deep Learning

80% match2024

Generalized Uncertainty of Deep Neural Networks: Taxonomy and Applications

80% match2023

On weight and variance uncertainty in neural networks for regression tasks

80% match2025

Cite this paper

@article{tagasovska2023retrospective,
  title  = {Retrospective Uncertainties for Deep Models using Vine Copulas},
  author = {Nataša Tagasovska and Firat Ozdemir and Axel Brando},
  year   = {2023},
  eprint = {2302.12606},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2302.12606v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2302.12606.svg)](https://codeofpaper.com/paper/2302.12606)

Discussion