A Practical Guide to Deep Learning - Richard Zemel

Explore a comprehensive seminar on deep learning fundamentals and practical applications. Delve into topics including perceptrons, neural network architectures, gradient descent, backpropagation, and object recognition. Examine the brain-inspired design of deep learning models, focusing on vision and somatosensory systems. Investigate advanced concepts such as local receptive fields, weight sharing, kernels, and edge detection. Learn about three-dimensional networks, pooling techniques, and regularization methods. Gain insights from Richard Zemel, a distinguished researcher from the University of Toronto, as he presents at the Institute for Advanced Study's Computer Science/Discrete Mathematics Seminar II.

Intro
Perceptrons
Brain inspiration
Layers
Classification
Gradient Descent
Back Propagation
Object Recognition
Vision and Vision
Neural networks
Local receptive fields
Somatosensory strip
Stride
Receptive fields
Feature detector
Weight sharing
Kernels
Edge detection
Threedimensional networks
Summary
pooling
upli
pooling layer
rotation layer
Natron
Regularizers