References: The handwritten-digit problem

Source material

Source curriculum (structural mirror, cited as further study):
• 3Blue1Brown, Neural Networks, Chapter 1: "But what is a Neural Network?"
  Creator: Grant Sanderson (text adaptation by Josh Pullen)
  Lesson page: https://www.3blue1brown.com/lessons/neural-networks
  Series index: https://www.3blue1brown.com/?topic=neural-networks
  License: copyright Grant Sanderson; videos published on his site and YouTube
Clawdemy's lessons are original prose that follows the pedagogical arc of this
series. We do not reproduce or transcribe the videos; we cite them as the
recommended companion. All rights to the original videos remain with the creator.

Watch this next

• But what is a Neural Network? (3Blue1Brown) by Grant Sanderson. The chapter this lesson mirrors, on the creator’s own site. This lesson stops at the motivating problem on purpose; the video keeps going and starts showing what is inside the network. About twenty minutes, and the animation is the reason this series is the standard recommendation for neural network intuition. Watch the opening now for the framing, then return to it as you move through lessons 2 to 4 of this track.

Going deeper

A short, durable list. Each link is a specific next step, not a generic pile.

• The MNIST database of handwritten digits, maintained by Yann LeCun and colleagues. This is the actual dataset behind the “handwritten digit” problem: 70,000 labeled 28x28 images that nearly every introduction to neural networks uses. If you want to see the real examples a system learns from, this is the source.

• Neural Networks and Deep Learning, Chapter 1 by Michael Nielsen. A free online book that opens, like this lesson, with handwritten digit recognition and then builds the full mechanism in careful prose. The natural deeper companion once you want the how behind the what.

Adjacent topics

Where this leads inside this track.

• Neurons as numbers, layers as structure (lesson 2). This lesson named the goal: a function from 784 numbers to 10. The next lesson opens that function up and shows it is built from layers of simple units, each holding a single number. Everything mechanical in the track starts there.

• What “learning” really means (lesson 5). We said the system “finds the pattern from examples” and deliberately did not explain how. Lesson 5 makes that precise: learning is adjusting the function’s internal numbers to make its answers on the labeled examples less wrong.