🕸 Neural networks: the basics

Unit 2 · How AI works · 10 min

A neural network has layers of "neurons". Each multiplies its inputs by weights, adds them and passes the sum through an activation function. In training, the network compares its prediction with the right answer and nudges the weights to reduce the error.

Words in this lesson

• 神经网络 (shén jīng wǎng luò) — neural network
• 权重 (quán zhòng) — weight
• 训练 (xùn liàn) — training

Project: Draw a small 3-layer network (input, hidden, output) and label each layer's job.