Ranking networks for fault diagnosis

Supervisor: Daira Pinto Prieto (d.pintoprieto@vu.nl)

Background

MWe talk about fault diagnosis in the context of cyber-physical systems. When a cyber-physical systems goes down, it is a big challenge to identify the root-cause of what is failing. Typically, companies offer diagnosis applications that assist the service engineers with identifying such root-cause.

In this context, some sort of reasoning must be encoded into the diagnosis application, aiming to aggregating the available information about the system and its symptoms and providing with a solution. This problem has been tackled in the literature by using Bayesian networks. This approach impose the strong assumptions of probability theory that can be relaxed by using other uncertainty representations, such as ranking functions. In this project, we want to explore the advantages of applying ranking networks to this problem, and compare them to Bayesian networks.

Description

This project aims to explore the application of possibility function networks to the domain of fault diagnosis. To this end, the student will work with other students on the creation of a small dataset for fault diagnosis and a Bayesian networks that will be used as a baseline model. Afterwards, the student will propose an end-to-end solution for fault diagnosis by applying ranking networks. Some milestones of this project are:

• Creating (collaboratively) a (small) fault diagnosis dataset.
• Creating (collaboratively) a Bayesian network for fault diagnosis.
• Getting familiar with ranking theory.
• Applying ranking networks to solve the fault diagnosis problem.
• Analyzing the obtained results.

Notice that this project has a collaborative part with other BSc students.

Literature

Rienstra, T. (2019, July). Ranked programming. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI-19) (pp. 5758–5764). International Joint Conferences on Artificial Intelligence Organization. [PDF]