Overview
Explore evolutionary testing simulation techniques for out-of-distribution images in this 56-minute technical talk from ANITI Toulouse. Delve into the SiMOOD (Simulation with Out Of Distribution) methodology, learning how to apply evolutionary algorithms to test and improve machine learning models' robustness against unfamiliar or unexpected inputs. Gain insights into creating diverse and challenging test scenarios, evaluating model performance under various conditions, and iteratively enhancing AI systems to handle real-world complexities and edge cases more effectively.
Syllabus
FOCUS TECHNIQUE ANITI SiMOOD Evolutionary Testing Simulation with Out Of Distribution Images
Taught by
ANITI Toulouse