`

Timezone: »

 
Poster
Single-Model Uncertainties for Deep Learning
Natasa Tagasovska · David Lopez-Paz

Wed Dec 11 10:45 AM -- 12:45 PM (PST) @ East Exhibition Hall B + C #51
in Algorithms -- Uncertainty Estimation »

We provide single-model estimates of aleatoric and epistemic uncertainty for deep neural networks. To estimate aleatoric uncertainty, we propose Simultaneous Quantile Regression (SQR), a loss function to learn all the conditional quantiles of a given target variable. These quantiles can be used to compute well-calibrated prediction intervals. To estimate epistemic uncertainty, we propose Orthonormal Certificates (OCs), a collection of diverse non-constant functions that map all training samples to zero. These certificates map out-of-distribution examples to non-zero values, signaling epistemic uncertainty. Our uncertainty estimators are computationally attractive, as they do not require ensembling or retraining deep models, and achieve state-of-the-art performance.

Keywords: Deep Learning · Algorithms · Uncertainty Estimation
[ Paper]  [ Poster] 

Author Information

Natasa Tagasovska (University of Lausanne)
David Lopez-Paz (Facebook AI Research)

More from the Same Authors