Master MVA / Responsible Machine Learning

2024/2025

*** Syllabus
Machine Learning is a fantastic scientific field but it is also involved in a (geo)political arena. And it may seriously harm individuals or societies.
The course tackles the three main issues of the implementation of Machine Learning from an ethical perspective: privacy, fairness, and explainability.
We introduce the main concepts and offer an overview of how such ethical considerations may be taken into account from a technical perspective.

***Instructor:
Prof. Nicolas VAYATIS

***Contact email :
<nicolas/-dot-/vayatis/-at-/ens-paris-saclay.fr>

***Course schedule and location:

Tuesdays 8:30am-10:30am at ENS Paris-Saclay
ENS Paris-Saclay (1Z25 or 1Z34)
Date Room Session Topics Material
Tuesday January 21
 1Z34 Lecture #1 Introducing responsible ML
Tuesday January 28 1Z25 Lecture #2 On Privacy-preserving Machine Learning Slides / Homework #1
Tuesday February 4 1Z34 Lecture #3 Machine Learning Fairness Slides / Homework #2
Tuesday February 11 1Z25 Lecture #4 Explainability in Machine Learning / Wrap-up
 Slides / Homework #3
Tuesday March 4
 1Z25 Evaluation
 Exam / Student talks

***Evaluation:

G1 = grade of the one-hour exam
G2 = grade of the ten-minute presentation
Final grade = (G1+G2)/2

***Readings

**On privacy

*Dwork, Roth (2014): <https://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf>
*Rigaki, Garcia (2023): <https://dl.acm.org/doi/pdf/10.1145/3624010>
*Kairouz et al. (2020): <https://arxiv.org/pdf/1912.04977>

**On fairness
Kleinberg, Mullainathan, Raghavan (2017): <https://doi.org/10.4230/LIPIcs.ITCS.2017.43>
Hardt, Price, Srebro (2016): <https://proceedings.neurips.cc/paper_files/paper/2016/file/9d2682367c3935defcb1f9e247a97c0d-Paper.pdf>
Mitchell et al. (2021): <https://cs.uwaterloo.ca/~jhoey/teaching/cogsci600/papers/Mitchell2021.pdf>
Domini et al. (2018): <https://arxiv.org/pdf/1802.08626>

  *************