Limits of Deep Learning
Tim, Florian, and Wieland discuss the limitations of deep learning models in understanding what they learn and the need for new objectives beyond just accuracy. They explore the challenges in optimizing for tasks and the impact of architecture, tasks, and optimizers on the solutions found by machines.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#040 - Adversarial Examples (Dr. Nicholas Carlini, Dr. Wieland Brendel, Florian Tramèr)
Related Questions