Continuous Time Evidential Distributions for Irregular Time Series

TL;DR

EDICT is a novel approach that models uncertainty in irregular time series using evidential distributions, enabling accurate and calibrated predictions at any time point, especially in sparse and noisy data scenarios.

Contribution

The paper introduces EDICT, a continuous-time evidential distribution method for irregular time series, improving uncertainty quantification and prediction flexibility.

Findings

EDICT achieves competitive classification performance.

It provides well-calibrated uncertainty estimates.

Enables uncertainty-guided inference in noisy data.

Abstract

Prevalent in many real-world settings such as healthcare, irregular time series are challenging to formulate predictions from. It is difficult to infer the value of a feature at any given time when observations are sporadic, as it could take on a range of values depending on when it was last observed. To characterize this uncertainty we present EDICT, a strategy that learns an evidential distribution over irregular time series in continuous time. This distribution enables well-calibrated and flexible inference of partially observed features at any time of interest, while expanding uncertainty temporally for sparse, irregular observations. We demonstrate that EDICT attains competitive performance on challenging time series classification tasks and enabling uncertainty-guided inference when encountering noisy data.