AINS6003 Deep Learning & Neural Networks#
This book is the instructional hub for AINS6003 Deep Learning & Neural Networks (Aurnova MSAI core course AIN6003).
It is organized as one unified MyST Jupyter Book so learners and instructors navigate a single artifact:
book prose for conceptual understanding
assignments for applied practice (Thebe-enabled notebooks)
slides for presentation flow (RISE-ready notebooks)
narration for asynchronous delivery
instructor notes for teaching guidance
notebooks for executable exploration
::{admonition} Work in GitHub Codespaces :class: tip
Launch a browser-based development environment for the full AINS6003 Deep Learning & Neural Networks repository:
Open this course in GitHub Codespaces
::{admonition} Use Google Colab for Notebook Labs :class: note
Many students in this program are new to Python. Colab is the recommended starting point for notebook labs because it opens in the browser and does not require installing Python, using a terminal, or understanding Git. Students can run the cells, change one small value, and write what they observe.
Start with Colab for labs. It gives students a notebook in the browser without requiring Git, a terminal, or local Python setup. Use Codespaces only when an instructor asks students to work with the full repository or a multi-file project.
Activity |
Best Environment |
Why |
|---|---|---|
Open and run a module lab |
Colab |
Easiest path for students who have never used Python. |
Change one parameter, threshold, feature, or model setting |
Colab |
The lab exercises are small, guided, and safe to rerun. |
Write short answers or observations from a lab |
Colab |
Students can submit a shared Colab link or exported notebook when allowed. |
Work on a larger graded project with several files |
Codespaces |
Keeps notebooks, scripts, data files, and instructions together. |
Use GitHub, terminal commands, tests, or repository history |
Codespaces |
Better for instructor-guided technical workflows, not first-time Python practice. |
Lab |
Launch |
|---|---|
Module 1 Lab |
|
Module 2 Lab |
|
Module 3 Lab |
|
Module 4 Lab |
|
Module 5 Lab |
|
Module 6 Lab |
|
Module 7 Lab |
|
Module 8 Lab |
Course arc (8 modules)#
From neurons to multilayer networks
Backpropagation and automatic differentiation
Optimization, loss, and regularization
Convolutional neural networks for vision
Sequence models: RNNs and LSTMs
Attention and transformers
Generative models and applications
GPU workflows, scale, and deployment
GPU-backed labs are scaffolded for Modules 3–8. Students can use the course Codespace or an approved cloud GPU environment where noted in the syllabus.