OAFP 6050 - Decoding biases in Artificial Intelligence
The course invites students to explore the issues of discrimination in AI. The deluge of public data and the advent of deep learning techniques in the last two decades have generated a lot of hope for designing more personalized and balanced public policies. However, using machine learning algorithms to make public decisions about every citizen has not come without generating fierce criticisms from activists and the general public.
In this class, we invite participants to get their hands on data and code and investigate on how and why public data and state-of-the-art algorithms, far from being neutral and objective, may inherently produce discrimination toward certain populations. Conversely, we will also use machine learning techniques to mine large datasets and try to characterize systematic race or gender biases.
The objectives of the class are three fold: (i) take some critical distance from the ongoing debates on the algorithmic fairness and its application to data-driven policy, (ii) learn and practice large data collection, manipulation and main families of machine learning algorithms, (iii) create your own research plan to investigate some original dataset or question in the collective project.
Jean-Philippe COINTET,Béatrice MAZOYER
● In Class Presence: 2 hours a week / 24 hours a semester
● Online learning activities: 20 minutes a week / 4 hours a semester
● Reading and Preparation for Class: 45 minutes a week / 18 hours a semester
● Research and Preparation for Group Work: 2 hours a week / 24 hours a semester
● Research and Writing for Individual Assessments: 20 minutes a week / 4 hours a semester
No coding ability is required. We are still expecting students to be curious about computer science, algorithms and their role in our world.
There will be two main assessments during the semester. The most important one (2/3 of the grade) is a collective project that will be presented during the last session of the semester. In this original project, students will be required to identify and collect a dataset online, design an experimental plan to analyze its inherent biases, and finally visualize and reflect upon the systematic discriminations embedded in the dataset. The final delivery will take the form of a website. An individual take-home paper will also be graded around week 7. Active participation during the class can also be rewarded with an extra point.
The pedagogical format is strongly oriented toward a workshop-style class. Typically, a short theoretical introduction will first be given. A discussion of the readings will follow before the class turns into applied mode, wherein students will practice data coding by themselves.