| MATH80630 | Lectures | Assignments | Project | Office hour

Trustworthy Machine Learning (MATH80630): Fall 2022

assingments.md

Lecture Schedule

---

1- Week 1 (September 2): Class introduction and machine learning review

• Lecture: Slides

---

2- Week 2 (September 9): Statistical Fairness Definitions in Machine Learning

• Lecture: Slides
• Reading:
  • Fairness definitions explained
  • Measurement and Fairness
  • Fairness Through Awareness
  • Interventions over Predictions:Reframing the Ethical Debate for Actuarial Risk Assessment

---

3- Week 3 (September 16): Causal Fairness Definitions in Machine Learning

• Lecture: Slides
• Reading:
  • Causal fairness analysis

---

4- Week 4 (September 23): Sources of Data Biases and Pre-Processing Fairness Approaches

• Lecture: Slides
• Reading:
  • Data preprocessing techniques for classificationwithout discrimination
  • Unbiased look at dataset bias
  • Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries
  • Algorithmic Bias: From Discrimination Discovery to Fairness-aware Data Mining

---

5- Week 5 (September 30): Fair Representation learning and In-Processing Fairness techniques

• Lecture: Slides
• Reading:
  • Learning Fair Representations
  • Learning adversarially fair and transferable representations
  • Fairness-aware Learning through Regularization Approach

---

6- Week 6 (October 7): Post-processing Fairness techniques, Intro to Privacy-Preserving Machine Learning and Differential Privacy

• Lecture: Slides
• Reading:
• • Equality of Opportunity in Supervised Learning
  • The Algorithmic Foundations of Differential Privacy, Chapters 1-3, Dwork, C., and Roth, A.
  • Robust De-anonymization of Large Sparse Datasets

---

7- Week 7 (October 14): Applications of DP in ML

• Lecture: Slides
• Reading:
  • Differentially private empirical risk minimization
  • What Can We Learn Privately?, Kasiviswanathan, S. et al.
  • Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data (optional)

---

8- Week 8 (October 21): Reading Week

• No Class

---

9- Week 9 (October 28): DP in ML optimization, DP-SGD, and Federated Learning

• Lecture: Slides
• Reading:
  • Deep Learning with Differential Privacy
  • FedAVG: Communication-Efficient Learning of Deep Networks from Decentralized Data
  • Advances and Open Problems in Federated Learning

---

10- Week 10 (November 4): Federated Optimization, Privacy and Fairness in Federated learning, Intersection of Privacy and Fairness

• Lecture: Slides
• Reading:
  • Advances and Open Problems in Federated Learning
  • Federated optimization in heterogeneous networks
  • Fair Resource Allocation in Federated Learning
  • Practical Secure Aggregationfor Privacy-Preserving Machine Learning (optional)

---

11- Week 11 (November 11): Robustness in ML, adverserial attacks, and FL and robustness

• Lecture: Slides

• Reading:

  • Practical Black-Box Attacks against Machine Learning
  • Intriguing properties of neural networks
  • Certified Adversarial Robustness via Randomized Smoothing
  • Adversarial machine learning at scale

---

12- Week 12 (November 18): Intro to Model understanding, Interpretability vs. Explainability, and Post-hoc Explanations: local approaches

• Lecture: Slides
• Reading:
  • Towards a rigorous science of interpretable machine learning
  • Human-in-the-loop Models & Explanations: Faithful and Customizable Explanations of Black Box Models
  • Understanding Black-box Predictions via Influence Functions
  • Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead.
  • Manipulating and Measuring Model Interpretability

---

13- Week 13 (November 26): Post-hoc Explanations: local vs. global approaches

• Lecture: Slides
• Reading:
  • Sanity Checks for Saliency Maps
  • A Unified Approach to Interpreting Model Predictions
  • “Why Should I Trust You?”: Explaining the Predictions of Any Classifier
  • Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)

---

14- Week 14 (December 2): Project Preperation and Neurips

• No Class

---

15- Week 15 (December 9): Project Presentation

• Room: Bleu, Salle Rona (Côte-Sainte-Catherine 1er étage, capacité 30)