Timezone: »

Quantification of Uncertainty with Adversarial Models
Kajetan Schweighofer · Lukas Aichberger · Mykyta Ielanskyi · Günter Klambauer · Sepp Hochreiter

Tue Dec 12 08:45 AM -- 10:45 AM (PST) @ Great Hall & Hall B1+B2 #1211
Event URL: https://github.com/ml-jku/quam »

Quantifying uncertainty is important for actionable predictions in real-world applications. A crucial part of predictive uncertainty quantification is the estimation of epistemic uncertainty, which is defined as an integral of the product between a divergence function and the posterior. Current methods such as Deep Ensembles or MC dropout underperform at estimating the epistemic uncertainty, since they primarily consider the posterior when sampling models. We suggest Quantification of Uncertainty with Adversarial Models (QUAM) to better estimate the epistemic uncertainty. QUAM identifies regions where the whole product under the integral is large, not just the posterior. Consequently, QUAM has lower approximation error of the epistemic uncertainty compared to previous methods. Models for which the product is large correspond to adversarial models (not adversarial examples!). Adversarial models have both a high posterior as well as a high divergence between their predictions and that of a reference model. Our experiments show that QUAM excels in capturing epistemic uncertainty for deep learning models and outperforms previous methods on challenging tasks in the vision domain.

Author Information

Kajetan Schweighofer (ELLIS Unit / University Linz)
Lukas Aichberger (ELLIS Unit / University Linz)
Günter Klambauer (Johannes Kepler University Linz)
Sepp Hochreiter (ELLIS Unit / University Linz)

More from the Same Authors