MIT 6.S191: Introduction to Deep Learning
This annual MIT course provides an introduction to deep learning methods with an emphasis on neural networks, including convolutional, recurrent, and generative adversarial networks. It covers foundational theory and practical applications with TensorFlow. Lecture videos, slides, and labs are publicly available.
More resources on Neural Networks
ImageNet Classification with Deep Convolutional Neural Networks
This paper introduced AlexNet, a groundbreaking convolutional neural network that significantly outperformed previous methods in the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) 2012. It marked a turning point for deep learning's resurgence and practical application.
Deep Learning
A seminal review paper by three pioneers of deep learning, providing an overview of the field, its history, key concepts, and future directions. It's an excellent read for understanding the landscape of deep learning.
StatQuest with Josh Starmer - Neural Networks videos
Josh Starmer's StatQuest provides clear, concise, and often humorous explanations of statistical and machine learning concepts. His videos on neural networks break down complex ideas into easily digestible 'quests.'
3Blue1Brown - Neural Networks series
Grant Sanderson's 3Blue1Brown channel offers visually intuitive and mathematically rigorous explanations of complex topics. His series on neural networks is particularly praised for making the core concepts, like backpropagation, understandable through animated visuals.
Neural Networks and Deep Learning
A foundational and widely acclaimed online book that explains neural networks and deep learning concepts from scratch, with clear explanations and interactive examples. It's an excellent resource for building a strong theoretical understanding.
Andrew Ng's Deep Learning Specialization
This specialization is one of the most popular and comprehensive introductions to deep learning. It covers foundational concepts of neural networks, including how to build and train them, and delves into convolutional networks, recurrent networks, and more. While the certificate requires payment, auditing the courses (watching lectures, accessing many materials) is typically free.