I Find Your Lack of Uncertainty in Computer Vision Disturbing
This is an incremental meta-analysis highlighting a critical safety issue for real-world computer vision applications involving human lives.
The paper identifies a widespread lack of proper epistemic uncertainty quantification in computer vision models, which can lead to unreliable high-stakes decisions, and urges the community to adopt calibrated uncertainty methods for out-of-distribution detection.
Neural networks are used for many real world applications, but often they have problems estimating their own confidence. This is particularly problematic for computer vision applications aimed at making high stakes decisions with humans and their lives. In this paper we make a meta-analysis of the literature, showing that most if not all computer vision applications do not use proper epistemic uncertainty quantification, which means that these models ignore their own limitations. We describe the consequences of using models without proper uncertainty quantification, and motivate the community to adopt versions of the models they use that have proper calibrated epistemic uncertainty, in order to enable out of distribution detection. We close the paper with a summary of challenges on estimating uncertainty for computer vision applications and recommendations.