← Back to Courses
AI & Machine Learningadvanced
AI-301

Deep Learning & Neural Networks

When you finish, you can

Build and train neural networks for images, sequences and text, and explain how transformers and generative models actually work.

Start hereYou need AI-201 and comfort with Python. Use free GPU time on Colab; nothing here needs a paid account.

9 hours total5 Chapters20 Lessons1 ExamsFREE
Start This CourseFree. No account required.

Updated · What changed

How this course updates itself

Every day we scan the latest information in every field. We share with you what we think is a change as we go through regular course curriculum work, so it's the most current thing possible.

Nothing in the news has changed what this course teaches. The daily scan last checked it on ; when something does change, a dated note with sources appears here.

Debate a counselor

Take a position on Deep Learning & Neural Networks. The AI counselor argues the other side, asks you questions back, and cites a source for every fact it states. Sources listed under a reply are the pages it actually found; if none are listed, it is arguing from reasoning only. It is an AI and it can be wrong: check the sources.

    Discuss this course

    A room for people taking this course, in any language, translated for every reader. Open breakout rooms for one question at a time.

    Enter the course room
    Prerequisites::AI-201 Machine Learning

    Course Syllabus

    9h total

    Practice

    1. Build a two-layer network on the MNIST digits in a few lines of PyTorch or Keras.
    2. Train for five epochs; plot training and validation loss.
    3. Add dropout and retrain; compare the two curves.

    Check

    Point to the epoch where overfitting begins on your plot and explain what regularization did to it.

    Topics Covered

    deep-learningneural-networkstransformers

    Ready to start?

    Free. No account required. Progress saves in this browser — no account exists.

    Start This Course