Faults in deep reinforcement learning programs: a taxonomy and a detection approach.
Amin NikanjamMohammad Mehdi MorovatiFoutse KhomhHoussem Ben BraiekPublished in: Autom. Softw. Eng. (2022)
Keyphrases
- action space
- reinforcement learning
- state space
- markov decision processes
- false alarms
- detection algorithm
- detection accuracy
- detection method
- function approximation
- detection rate
- multi agent
- automatic detection
- learning algorithm
- computer programs
- fault diagnosis
- object detection
- software engineering
- event detection
- hierarchical structure
- false positives