Matrices
A matrix is a table of numbers. Neural networks use matrices to turn an input into an answer. In this lesson you read a matrix, multiply it with a vector, and see how one layer of a network works.
New words
- Matrix
- A table of numbers with rows and columns.
- Shape
- The size of a matrix: rows × columns.
- Matrix multiplication
- Take dot products of rows with columns. NumPy writes it as @.
- Transpose
- Flip a matrix so rows become columns. NumPy writes it as .T.
- Weights
- The numbers a model learns. A layer stores them in a matrix.
- Bias
- A number added after the multiplication to shift the answer.
What you will learn
- Where do matrices show up in AI?
- What is a matrix?
- How do you multiply a matrix with a vector?
- What is the shape rule?
- Matrix times matrix, and the transpose
- What is a neural network layer?
- Is a photo a matrix?
- Try it: multiply row by row
- What mistakes do people make?
- Explain it to a friend
- Key takeaways
10 quiz questions at the end.
The full lesson, with interactive demos and the quiz, opens after you sign in with Google. It is free.