Practical Applications of ML in Software Testing
summary
Case study classifies bug severity from bug summaries using supervised learning with TF‑IDF NLP features, trained on roughly 400 real bugs split for train/test. Initial model shows a 73% overall accuracy with a bias toward severity 2; merging classes (2 vs. others) raises performance to 82% accuracy. Practical examples include automatic API/UI test generation from observed request/response schemas, visual element recognition via test.ai visual recognition, execution anomaly detection by duration, and clustering to surface recurring error groups.
AI-written summary · about
chapters (23)
- 0:00Intro
- 1:25AGENDA
- 2:47INTRODUCTION
- 4:16Me vs. Challenges
- 4:32Can They Help?
- 7:38Already being used widely
- 9:29Learning: Meaning
- 14:42Background
- 17:56Algorithms
- 18:30Applications
- 18:41Software Testing Stages
- 20:24Definition / Case Generation
- 26:33Implementation / Test Generation
- 27:57Implementation / Code Generation
- 31:48Implementation / Execution
- 34:06Maintenance / Refactoring
- 35:56Maintenance / Self Healing
- 36:50Maintenance / Prioritization
- 38:23Maintenance / Bug Management
- 42:35Bug Severity Estimation
- 44:35Who is right?
- 46:20Clustering
- 48:11Summary
related talks
25mFuture of Software Testing: Machine Learning Assistance
44mThe Future is Today: Leveraging AI in Software Testing
46mSW Testing: Can ML save us?
16mTransforming Test Automation with AI and Machine Learning
01h 04mWhat Testers Need from Machine Learning and Artificial Intelligence
40mThe Applicability of Machine Learning and Neural Networks in Test Automation
48mUsing Machine Learning to find Linux bugs
40mFrom Bugs to Brilliance: Leveraging AI in Modern Software Testing