Code Examples

This section collects Jupyter notebooks with Python implementations of the adversarial attacks and defenses covered in the lectures. The notebooks accompany the lecture slides and serve as starting points for the homework assignments and the course project. Students are encouraged to use the GPUs provided by Google Colab for running the notebooks.

Each notebook can be viewed on this website, opened directly in Google Colab, or downloaded from the GitHub repository.

Adversarial Attacks and Defenses

Notebook	Open in Colab
FGSM Attack in TensorFlow	Colab
White-box Evasion Attacks in TensorFlow	Colab
White-box Evasion Attacks in PyTorch	Colab
Black-box Boundary Attack	Colab
LLM Supervised Fine-tuning	Colab
Jailbreak Attacks on LLMs	Colab
Adversarial Training	Colab
Attack on Network Intrusion Detection	Colab

Hello World of ML

Introductory examples of training and evaluating machine learning models, from classical models on tabular data to fine-tuning large language models and vision-language models. A one-page overview is available as a PDF.

Year	Notebook	Open in Colab
2005	Logistic Regression on Iris	Colab
2010	Random Forest on Breast Cancer	Colab
2014	XGBoost on Titanic	Colab
2015	MLPs on MNIST	Colab
2018	ResNet on CIFAR-10	Colab
2021	DistilBERT on IMDb	Colab
2023	Llama 2 on Alpaca	Colab
2025	Qwen3 GRPO on MATH-500	Colab
2026	VLM Reasoning and Tool Use on MATH-500	Colab