Jakob Smedegaard's lectures
Jakob Smedegaard Andersen · COLING 2020
1149 - Word-Level Uncertainty Estimation for Black-Box Text Classifiers using RNNs
Common uncertainty estimation approaches remain as black-boxes without explaining which features have led to the uncertainty of a prediction. We introduce an approach to decompose and visualize the uncertainty of text classifiers at the level of words. We aim to enable a deeper inspection and reasoning about unreliable model behaviours by providing detailed explanations of uncertainties.