Evidential Turing Processes

TL;DR

This paper introduces a novel probabilistic classifier that combines Evidential Deep Learning, Neural Processes, and Neural Turing Machines to achieve reliable uncertainty estimation, calibration, and out-of-domain detection in classification tasks.

Contribution

It presents an original unified model capable of total uncertainty quantification, excelling in calibration, out-of-domain detection, and predictive reliability with a single predictor.

Findings

Outperforms existing methods in calibration and out-of-domain detection

Achieves reliable predictive uncertainties across five classification tasks

Provides an efficient and implementation-friendly approach for safety-critical applications

Abstract

A probabilistic classifier with reliable predictive uncertainties i) fits successfully to the target domain data, ii) provides calibrated class probabilities in difficult regions of the target domain (e.g.\ class overlap), and iii) accurately identifies queries coming out of the target domain and rejects them. We introduce an original combination of Evidential Deep Learning, Neural Processes, and Neural Turing Machines capable of providing all three essential properties mentioned above for total uncertainty quantification. We observe our method on five classification tasks to be the only one that can excel all three aspects of total calibration with a single standalone predictor. Our unified solution delivers an implementation-friendly and compute efficient recipe for safety clearance and provides intellectual economy to an investigation of algorithmic roots of epistemic awareness in deep neural nets.