Skip to content
Clawdemy

References: Seeing the field whole

Source material

Section titled “Source material”
Source curriculum (structural mirror, cited as further study):
• MIT 6.S191, "Introduction to Deep Learning" (full course)
  Instructors: Alexander Amini and Ava Amini (MIT)
  Course page: https://introtodeeplearning.com
  Code and labs: https://github.com/aamini/introtodeeplearning
  License: MIT (slides, code, and labs); videos are YouTube standard
  Required attribution: "© Alexander Amini and Ava Amini, MIT 6.S191:
    Introduction to Deep Learning, IntroToDeepLearning.com"
This lesson is an original synthesis; it mirrors no single lecture. The track as
a whole follows the pedagogical arc of MIT 6.S191. Clawdemy's lessons are original
prose; we do not reproduce or transcribe the lectures, and cite them as the
recommended companion. Course materials are used under their MIT license with the
attribution above; all rights to the original videos remain with the creators.

Watch this next

Section titled “Watch this next”
  • MIT 6.S191: Introduction to Deep Learning (full course) by Alexander and Ava Amini. You now have the foundation to watch the whole course end to end and follow every lecture. It is refreshed every year, so it is also the best way to see the most current version of this survey, with the latest frontier material the field keeps adding. If you watched lectures as you went, this is the moment to take in the course whole.

Where to go next in Clawdemy

Section titled “Where to go next in Clawdemy”

The honest routing, by what you want to do.

  • Track 5 (Transformers and LLMs). Go here to understand modern language models in depth. This track covered attention and transformers “in brief”; Track 5 builds the full mechanics and how they scale into the assistants you use daily.

  • Track 13 (Build Neural Networks from Scratch). Go here to build what you just surveyed, in code, from first principles. The most convincing way to make the survey concrete.

  • Neural Network Intuition (the previous track). The engine this whole survey assumed, layers, weights, gradient descent, backpropagation. If any moment here rested on machinery you were unsure of, this is the foundation to firm up.

Adjacent topics

Section titled “Adjacent topics”

The arc of this track, for quick reference.

  • The unlock (lesson 1): depth, data, and compute arriving together.
  • The four shapes (lessons 2 to 8): sequences (recurrence, attention), images (convolution), generation (VAE, GAN, diffusion), decisions (reinforcement learning), all one engine, wired differently.
  • The limits (lesson 9): data hunger, brittleness, data-slant, and the lack of guarantees, the bounds that make the map trustworthy.