Towards Knowledge Uncertainty Estimation for Open Set Recognition

Teacher-Explorer-Student Learning: A Novel Learning Method for Open Set Recognition

When an unknown example, one that was not seen during training, appears, most recognition systems usually produce overgeneralized results and determine that the example belongs to one of the known classes. To address this problem, teacher-explorer-student (T/E/S) learning, which adopts the concept of open set recognition (OSR) to reject unknown samples while minimizing the loss of classification performance on known samples, is proposed in this study. In this novel learning method, the overgeneralization of deep-learning classifiers is significantly reduced by exploring various possibilities for unknowns. The teacher network extracts hints about unknowns by distilling the pretrained knowledge about knowns and delivers this distilled knowledge to the student network. After learning the distilled knowledge, the student network shares its learned information with the explorer network. Next, the explorer network shares its exploration results by generating unknown-like samples and feeding those samples to the student network. As this alternating learning process is repeated, the student network experiences a variety of synthetic unknowns, reducing overgeneralization. The results of extensive experiments show that each component proposed in this article significantly contributes to improving OSR performance. It is found that the proposed T/E/S learning method outperforms current state-of-the-art methods.

Uncertainty-Based Rejection in Machine Learning: Implications for Model Development and Interpretability

Uncertainty is present in every single prediction of Machine Learning (ML) models. Uncertainty Quantification (UQ) is arguably relevant, in particular for safety-critical applications. Prior research focused on the development of methods to quantify uncertainty; however, less attention has been given to how to leverage the knowledge of uncertainty in the process of model development. This work focused on applying UQ into practice, closing the gap of its utility in the ML pipeline and giving insights into how UQ is used to improve model development and its interpretability. We identified three main research questions: (1) How can UQ contribute to choosing the most suitable model for a given classification task? (2) Can UQ be used to combine different models in a principled manner? (3) Can visualization techniques improve UQ’s interpretability? These questions are answered by applying several methods to quantify uncertainty in both a simulated dataset and a real-world dataset of Human Activity Recognition (HAR). Our results showed that uncertainty quantification can increase model robustness and interpretability.