The field of prognostics and health management is moving towards a more integrated approach, combining risk assessment and fault prediction to provide a more comprehensive understanding of system reliability. This is achieved through the use of advanced modeling frameworks, such as continuous-time Bayesian networks, and techniques like graph convolutional networks and representation learning. The application of these methods enables the development of more accurate and robust fault diagnosis and prediction systems, which can be used in a variety of industrial domains. Noteworthy papers in this area include:

• A study on hierarchical knowledge guided fault intensity diagnosis, which proposes a novel framework for capturing dependencies among target classes and achieves superior results on real-world datasets.
• A paper on rethinking the probability of failure in mitigation safety functions, which argues that the traditional approach is not appropriate and proposes an alternative method leveraging the probability density function and expected degree of failure.