Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore multi-objective automated machine learning techniques that balance accuracy with robustness in this 47-minute seminar presentation. Learn how AutoML can be extended beyond traditional accuracy optimization to create trustworthy models that withstand various types of domain shifts and input perturbations. Discover research findings from the Horizon TAILOR project that demonstrate methods for developing neural networks capable of meeting multiple requirements simultaneously, including deployment constraints for satellite hardware and resilience against adversarial attacks. Examine specific hyperparameter tuning strategies that enhance model robustness while maintaining performance, and understand how these multi-objective approaches enable data scientists to focus on higher-level tasks by automating the complex trade-offs between accuracy, size, and trustworthiness in machine learning model selection and optimization.
