Skip to content

Applications

 

eXplainable Deep Neural Networks (xDNN)

 

 

xDNN is a prototype-based network that uses data density as its core mechanism.

Prototypes are selected data samples that the users can easily view, understand and analyze their similarity to other data samples.

Related papers: 

Angelov, P., & Soares, E. (2020). Towards explainable deep neural networks (xDNN). Neural Networks, 130, 185-194.

Soares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In 2019 International Joint Conference on Neural Networks (IJCNN) (pp. 1-8). IEEE.

Git:

https://github.com/Plamen-Eduardo/xDNN—Python

xDNN for model interpretability in the context of DRL

 

xDNN for object detection

xDNN is applied for the object detection context. Due to its transparent design, it allows an explainable error analysis which is crucial for high stake applications. 

xClass for Novelty detection

 

xClass extends xDNN and offers:

  • Real-time “novelty detection” through the recursive density and the m-sigma rule.
  • Weakly supervised structure that allows few 1 class and few data samples initially.
  • Automatic creation of new classes.
  • Transparent structure that allows humans to audit it.

Related papers: 

Angelov, P., & Soares, E. (2021). Detecting and learning from unknown by extremely weak supervision: exploratory classifier (xClass). Neural Computing and Applications, 1-13.

Soares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In 2019 International Joint Conference on Neural Networks (IJCNN) (pp. 1-8). IEEE.