Mozilla uses machine learning to fix bugs in Firefox

Mozilla is going to use software that should reduce the number of bugs in Firefox via machine learning. The so-called coding assistant is called Clever-Commit and was developed by game maker Ubisoft.

Clever-Commit combines data from Firefox’s bug tracking and version control system to discover patterns in the history of programming errors using artificial intelligence. According to Mozilla, this will ensure that the next versions of Firefox will be faster and more stable.

The company will use the software during the Firefox coding, testing and release process. By integrating Clever Commit throughout the development process, Mozilla believes it will eventually be able to intercept 60 to 80 percent of all bugs before they end up in the code. The developer expects to save hundreds of hours of risk analysis and bug detection in this way. Clever-Commit was developed by Ubisoft La Forge, the R&D division of game maker Ubisoft. Mozilla will contribute to the further development of the coding assistant, among other things by making its knowledge of programming languages ​​and bug tracking systems available.

Firefox releases a new version every six to eight weeks. According to Mozilla, such a new browser contains an average of eight thousand changes in the code, which includes JavaScript, C++ and Rust.