Section 00 of Tensors for Machine Learning. Load every dataset and confirm your runtime works before the workshop starts. Four words carry the rest, and you need no technical background for them.

Study here, then run the real thing: [notebook 00 on Colab](https://colab.research.google.com/g

Part I of Tensors for Machine Learning. A tensor is a way to organize numbers using one or more directions, called axes. This section teaches you to read those axes before doing anything more advanced.

If you are running late, cut appendix material — not Part I. Every later section uses t

Part II of Tensors for Machine Learning. The goal is simple:

Do not read a tensor as a list of numbers. Read every axis as a question: “What does this axis count?”

Two tensors can have identical shapes and mean nothing like each other. Chapter 2’s notation has no concept of “order ma

Part III, Block 1 of Tensors for Machine Learning. Two practical skills: indexing — choosing exactly the rows, columns or values you want — and broadcasting — applying a smaller set of numbers across a larger array without copying them by hand.

This is also where the workshop’s centra

Part III, Block 2 of Tensors for Machine Learning. One idea that prevents many silent bugs:

Changing a tensor’s shape is not the same as moving its axes.

Microscopes and cameras order their axes according to the hardware, not according to what a model expects. Getting this wrong does

Part III group exercise of Tensors for Machine Learning. A video pipeline is the path from a video file to the tensor a model actually receives:

raw file → decoded frames → preprocessed batch → model input → model output

The key idea:

**Every pipeline decision chooses what i

Part IV, Block 3 of Tensors for Machine Learning. One rule that looks compact in code and has a very simple meaning:

If an index disappears after ->, NumPy sums over it. If the index remains, it survives in the output.

Recommendation and search systems rank items by the dot product

Part IV, Block 4 of Tensors for Machine Learning. Use one question throughout:

Does an exact inverse exist? If not, what useful answer does the pseudoinverse give us instead?

In machine learning A is almost never square: one row per example, one column per feature, and far more exa

Part IV demo of Tensors for Machine Learning. One simple pattern:

Use the current state to create the next state, then repeat.

The same pattern appears in Fibonacci numbers, in power iteration, and in a real monthly airline-passenger forecast. The last of those combines recursion wit

Part IV, Block 5 of Tensors for Machine Learning. Section 01 drew the map of factorizations. This section walks it, on real data, and answers the two questions a practitioner actually has: which one do I reach for on this data, and what does it cost me?

The organizing idea is that all six

Part IV, Block 6 of Tensors for Machine Learning. One practical question:

How can we compress a tensor while keeping the meaning of its different axes?

PCA compresses a matrix — two axes. Real data often has more. Tucker decomposition generalizes PCA to a tensor of any order:

Part IV, Block 7 of Tensors for Machine Learning. Section 09 asked which factorization, and what does it cost one order down, on matrices. This section asks it of tensors.

A tensor decomposition is not only a compression technique. Each one makes a different assumption about **which struc

The close of Tensors for Machine Learning. This section has two jobs: connect the ideas from all thirteen sections, and hand you the extensions.

One idea connects sections 07, 09 and 10: when a problem has no exact answer and no true inverse, you do not give up. **You find the best stable