Abstract
本文:
Task: Review on the use of LLMs in software testing
Method: 1. analyzes 52 relevant studies
1. Intro
2. Background
2.1 Large Language Model
2.2 Software Testing
3. Paper Selection and Review Schema
3.1 Survey Scope
3.2 Paper Collection Methodology
3.3 Collection Results
4. Analysis from Software Testing
4.1 Unit Test Case Generation
4.2 Test Oracle Generation
4.3 Test Input Generation
4.4 Bug Analysis
4.5 Debug
4.6 Program Repair
5 Analysis from LLM
5.1 LLM Models
5.2 Types of Prompt Engineering
5.3 Input of LLM
5.4 Incorporating X with LLM
6 Challenges and opportunities
6.1 Extending to more tasks and more phases
6.2 Serving other types of testing and software
6.3 More Solid Benchmarks and Rigorous Evaluations
6.4 Boosting LLM Performance
7 Conclusion
- CDeepFuzz Landscape Language Software Readingcdeepfuzz landscape language software differentiation cdeepfuzz automatic reading pre-trained cdeepfuzz natural reading combinatorial cdeepfuzz learning reading cross-language information polycruise cdeepfuzz state-of-the-art cdeepfuzz the reading cdeepfuzz networks reading testing landscape distillation innovative knowledge landscape 地形landscape系统ue4