Sensorimotor Contingency: Amir Zamir

This lecture by Amir Zamir covers the concept of sensorimotor contingency and its role in understanding intelligence without representation. The lecture delves into the sensory manifold, the learned degrees of freedom in organisms, and the connection between vision and action. It also explores the developmental perspective, the importance of exploration, long childhood, and incremental learning in shaping intelligence. The presentation discusses curriculum learning, the pentomino test, deep architectures, and the benefits of using a curriculum strategy in machine learning algorithms.