Add Jupyter notebook assignments for Deep Learning FS26

[Copilot]

Course had no assignment material — adds three structured Jupyter notebooks covering foundational deep learning topics, each split into three consistent parts.

Notebook structure (uniform across all assignments)

• Part 1 — Task Description: Problem statement, task breakdown, possible solution approaches, and ASCII-art plot hints showing expected output shapes
• Part 2 — Implementation: Guided code stubs with TODO markers for students to complete
• Part 3 — Experiments & Analysis: Comparative experiments, visualization tasks, and written reflection questions

Assignments added (assignments/)

Notebook	Topic	Key concepts
Assignment1_Perceptrons_MLP.ipynb	Perceptrons & MLPs	AND/XOR gates, NumPy backprop, PyTorch nn.Module, decision boundaries
Assignment2_CNN.ipynb	Convolutional Neural Networks	Manual 2D conv, CIFAR-10, BatchNorm, feature map hooks, confusion matrix
Assignment3_RNN_LSTM.ipynb	RNNs & LSTMs	Sine-wave forecasting, char-level LM, text generation, hidden state heatmap

Other

• Updated README.md with course overview and assignment table, including three options for running notebooks (quickstart script, manual setup, and cloud via Google Colab/Binder)
• Added start_jupyter.sh launcher script that auto-installs dependencies, starts the Jupyter server, and prints the server URL (including login token) to the terminal for use in a browser or IDE (VS Code, PyCharm)

Original prompt

i need Jupyter notebok Each assignment is provided as Jupyter notebook
They include three parts each
Part 1: Task Description
Description of the problem to be solved
Details of the tasks and possible solutions
Possible outcomes of plotting functionality
→ Provides hints how solutions could look like

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[sing-git]

@copilot enter URL of the running Jupyter Server when i click jupyter to run this

[Copilot]

@copilot enter URL of the running Jupyter Server when i click jupyter to run this

The Jupyter server URL is printed directly in your terminal when you start the server. Here's the quickest way (commit acfff0f):

Run from the repo root:

bash start_jupyter.sh

The terminal will print something like:

http://127.0.0.1:8888/tree?token=abc123def456...

Copy that full URL (including the ?token=... part) and use it:

• Browser → paste into the address bar
• VS Code → Command Palette (Ctrl+Shift+P) → "Jupyter: Specify Jupyter Server for Connections" → paste the URL
• PyCharm / DataSpell → Settings → Tools → Jupyter → Jupyter Servers → add server → paste the URL

The README now has full step-by-step instructions for all three options (script, manual, and cloud via Google Colab/Binder with no install needed).