Computational Inductive Biases of Spatiotemporal Artificial Neural Networks

This lecture from the Fields Lab for Network Science features T. Anderson Keller from the Kempner Institute discussing the computational inductive biases present in spatiotemporal artificial neural networks. Explore how these biases influence network learning and performance across various spatiotemporal tasks. Delivered on April 4, 2025, the 59-minute presentation examines fundamental principles that shape how neural networks process and interpret information with spatial and temporal dimensions.