Estimating design quality of digital systems via machine learning

Although the term design quality of digital systems can be assessed from many aspects, the distribution and density of bugs are two decisive factors. This paper presents the application of machine learning techniques to model the relationship between specified metrics of high-level design and its associated bug information. By employing the project repository (i.e., high level design and bug repository), the resultant models can be used to estimate the quality of associated designs, which is very beneficial for design, verification and even maintenance processes of digital systems. A real industrial microprocessor is employed to validate our approach. We hope that our work can shed some light on the application of software techniques to help improve the reliability of various digital designs.