- author = "Matthew Leeke",
- booktitle = "2021 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech)",
- title = "Reducing Model Complexity and Cost in the Generation of Efficient Error Detection Mechanisms",
- year = "2021",
- pages = "26--34",
- abstract = "The design and location of error detection mechanisms (EDMs) is fundamental to the design of a dependable software system. The application of machine learning algorithms to fault injection data has been shown to be an effective approach for the generation of efficient EDMs. However, the complexity of the generated models and initial cost of generation represent barriers to the adoption of the approach. Addressing these challenges directly, this paper demonstrates that genetic programming can be used as an approach to reduce the complexity of the models generated and obviate the computational cost associated with the sampling and refinement stages of EDM generation. More specifically, it is shown that (i) genetic programming can be used to project the instance space of fault injection data sets into a space more amenable to learning, (ii) machine learning algorithms can be applied to the resultant projection to permit the generation of efficient EDMs with reduced model complexity, and (iii) the cost of generating efficient EDMs can be reduced by the approach because it obviates the need for data set sampling methods and model refinement.",
- keywords = "genetic algorithms, genetic programming",
-
DOI = "
doi:10.1109/DASC-PICom-CBDCom-CyberSciTech52372.2021.00020",
- month = oct,
-
notes = "Also known as \cite{9730167}
See also Phd: http://webcat.warwick.ac.uk/record=b2581851~S1 http://wrap.warwick.ac.uk/52394/1/WRAP_THESIS_Leeke_2011.pdf",
