Neural Network Extrapolation
Jan and Petar discuss the extrapolative nature of neural networks, exploring concepts like convex hulls and out-of-distribution generalization. They delve into the challenges of training models to perform well in unseen conditions, highlighting the complexities of neural network behavior beyond conventional bounds.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#85 Dr. Petar Veličković (Deepmind) - Categories, Graphs, Reasoning [NEURIPS22 UNPLUGGED]
Related Questions