Better C++ using Machine Learning on Large Projects
summary
Machine learning scores commits at pre-commit time using ~40 features (file age, lines added, author experience) and flags high-risk changes; the model reports about 70% precision and 71% recall and treats only ~5% of commits as >50% risk. For flagged commits the system builds partial abstract syntax trees, matches them to SZZ-labeled past bug-introducing commits across similar projects, and replays the AST edits to propose concrete fixes; risk prediction is live, automated patch suggestions are experimental.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:49Outline
- 10:08Machine Learning Features
- 14:12Machine Learning Performances
- 19:48Rainbow Six Siege
- 26:15Potential
related talks
47mBug prediction based on your code history
38mApplying machine learning to software development to reduce bugs
37mMachine Learning on Source Code
48mUsing Machine Learning to find Linux bugs
01h 31mClean(er) Code for Large Scale Legacy Applications
25mMachine Learning for source code analysis
53mBug prediction based on your code history
27mMachine Learning Practical Example