Human-Centric Machine Learning seminar (Winter 2022)

The Human-Centric Machine Learning seminar focuses on reading and presenting recent papers in human-centric machine learning, with a particular emphasis in societal aspects of machine learning (fairness, accountability, transparency), human-ai collaboration, strategic behavior and causality. The papers will cover both theoretical foundations as well as applications and empirical studies.

The seminar is open both for Bachelor and Master students.

As a student

You will read a few papers in one topic in human-centric machine learning and present an overview of the topic. There will be several possible topics to choose from (see below). To help you with the presentation, an instructor will be assigned to you to supervise you and provide advice for your presentation.

The presentations will take place on 3-4 different dates throughout the Winter semester. As a student, you will need to attend all presentations and ask questions to the presenters.

Important dates:

• Kickoff meeting: 26.10.22
• Topic selection: 02.12.22

Note: If you have any questions about the seminar, feel free to drop an email to Nastaran ([email protected]), Eleni ([email protected]) or Stratis ([email protected]).

Available topics

The available topics are the following:

• Interpretable Machine Learning
• Bias and Fairness
• Strategic Machine Learning
• Causality in Machine Learning
• Distribution-Free Uncertainty Quantification
• Human Machine Collaboration

A list of suggested papers for each topic can be found here: https://docs.google.com/document/d/1o7Rx9pbsMhn9ErR8n2owkIV6KALG9QVKgYqXjyHcLUg/edit?usp=sharing

When and where

Schedule and location:

• Room: Building G26, Room 112.
• Date and time: Three days during the Winter Semester. It will be announced beforehand.
• Kickoff meeting: Wednesday 26.10.22 15:00 - 16:00, Building G26, Room 112.
• Registration: send an email to [email protected] with your name and matriculation number.
• Topic selection: send an email to [email protected] with your preferred ranking for all the available topics.

Instructors

• Dr. Manuel Gomez-Rodriguez
• Nastaran Okati
• Eleni Straitouri
• Stratis Tsirtsis

Planned presentations

To be scheduled.