A Gentle Introduction to Neural Networks - with Python
Explore a gentle introduction to neural networks and learn how to create your own using Python in this 45-minute conference talk from EuroPython 2016. Gain an intuitive understanding of neural networks, their functionality, and their applications, regardless of your background in mathematics, computer science, or Python. Discover the search for AI, learn about biologically inspired neurons and networks, and understand the breakthrough of back propagation in training neural networks. Delve into Python's simplicity and ubiquity, explore Python notebooks, and work with the MNIST dataset. Implement a concise and efficient neural network class using numpy for matrix calculations, achieving 97.5% accuracy in recognizing handwritten numbers with just a few lines of code. Conclude with a Q&A session to address any remaining questions about neural networks and their implementation in Python.