An evidential classifier based on Dempster-Shafer theory and deep learning
This work addresses classification uncertainty for domains like image recognition and signal processing, but it is incremental as it integrates existing DS theory with deep learning.
The paper tackled the problem of set-valued classification by proposing an evidential classifier combining Dempster-Shafer theory with a convolutional neural network, resulting in improved classification accuracy and cautious decision-making by assigning confusing patterns to multi-class sets.
We propose a new classifier based on Dempster-Shafer (DS) theory and a convolutional neural network (CNN) architecture for set-valued classification. In this classifier, called the evidential deep-learning classifier, convolutional and pooling layers first extract high-dimensional features from input data. The features are then converted into mass functions and aggregated by Dempster's rule in a DS layer. Finally, an expected utility layer performs set-valued classification based on mass functions. We propose an end-to-end learning strategy for jointly updating the network parameters. Additionally, an approach for selecting partial multi-class acts is proposed. Experiments on image recognition, signal processing, and semantic-relationship classification tasks demonstrate that the proposed combination of deep CNN, DS layer, and expected utility layer makes it possible to improve classification accuracy and to make cautious decisions by assigning confusing patterns to multi-class sets.