Overview
Explore the construction of an analog neural network in vanilla Minecraft without mods or command blocks in this innovative video tutorial. Learn how Redstone wire power strengths carry signals through a hidden layer, incorporating nonlinearities, automatic backpropagation, and weight updates. Dive into explanations of Redstone components, analog multiplication techniques, and gradient descent for square root computation. Witness a live demonstration of the neural network in action, followed by a detailed breakdown of the network schema. Observe the network learning process as it adapts to a datapoint, and gain insights into the potential applications of this unique Minecraft creation.
Syllabus
- Intro & Overview
- Redstone Components Explained
- Analog Multiplication in Redstone
- Gradient Descent for Square Root Computation
- Neural Network Demonstration
- Network Schema Explained
- The Network Learns a Datapoint
- Outro & Conclusion
Taught by
Yannic Kilcher