Explore the history, motivation, and evolution of Deep Learning in this comprehensive lecture. Delve into the inspiration behind deep learning, tracing its roots and examining the history of pattern recognition. Learn about gradient descent and its computation through backpropagation. Discover the hierarchical representation of the visual cortex and follow the evolution of Convolutional Neural Networks (CNNs) from Fukushima to LeCun to AlexNet. Investigate practical applications of CNNs, including image segmentation, autonomous vehicles, and medical image analysis. Understand the hierarchical nature of deep networks and their advantageous attributes. Conclude by exploring the concepts of generating and learning features/representations in deep learning.
History, Motivation, and Evolution of Deep Learning
Overview
Syllabus
– Week 1 – Lecture
– Inspiration of Deep Learning and Its History, Supervised Learning
– History of Pattern Recognition and Introduction to Gradient Descent
– Computing Gradients by Backpropagation, Hierarchical Representation of the Visual Cortex
– Evolution of CNNs
– Deep Learning & Feature Extraction
– Learning Representations
Taught by
Alfredo Canziani
