Section for Cognitive Systems DTU Compute

02901 Advanced Topics in Machine Learning: Fairness and Machine Learning

August 26-30, 2019 at the Section for Cognitive Systems, DTU Compute

Description

As the development of machine learning algorithms advances rapidly, they play larger and larger roles for our societies. People both in- and outside the technical domain are rightfully concerned about the fairness of such algorithms:How can we explain, justify and assign accountability for automated decisions made by complex ML models? How do deep learning algorithms propagate existing bias and stereotypes? How can we define and measure algorithmic fairness?While those questions certainly aren't engineering problems alone, this year's summer school will consider them from a technical perspective.

The course consists of five days (Monday-Friday) of lectures and exercises. The lectures will cover theoretical aspects such as explainability of deep learning models, causal reasoning and fairness metrics. Technical aspects, programming and application of the developed concepts will be explored in tutorials and exercises. The course (2.5 ECTS points) is passed by handing in a small report on one of the topics covered in the course. For further course details click here.

Location

Technical University of Denmark, building 324, room 060.

Hours

Every day August August 26-30, 2019, exact starting and end times TBA.

Lecturers

Confirmed speakers:

• Thore Husfeldt, IT University Copenhagen: Introduction to Algorithmic Fairness.
• Natalie Schluter, IT University Copenhagen: Fair Natural Language Processing.
• Lars Kai Hansen, Technical University of Denmark: Explainabiliy and Fairness.
• Martin Mose Bentzen, Technical University of Denmark and
• Jens Ulrik Hansen, Roskilde University: A Causal Take on Algorithmic Fairness.
• Indré Žliobaité , University of Helsinki: Measuring algorithmic fairness.
• Silvia Chiappa, DeepMind: A Causal Take on 'Algorithmic Fairness III'.

Course programme

Requirements

General understanding of machine learning, statistical modeling, mathematics and computer science. Programming experience, ideally in Python. For the course you are required to bring your own laptop computer with a Python installation.

Registration

Registration is closed. Accepted participants have been notified. If you have any questions regarding the registration please email Wanja Andersen waan@dtu.dk

For academics (masters and PhD students) there is no registration fee for the course. For all other participants a course fee of DKK 8250 will be charged. Participants are to cover all other costs such as food, accommodation, and travel expenses.

Practical information

For practical information regarding transportation and accommodation click here.

2018 version of 02901 Advanced Topics in Machine Learning

2017 version of 02901 Advanced Topics in Machine Learning

2016 version of 02901 Advanced Topics in Machine Learning

2015 version of 02901 Advanced Topics in Machine Learning

2014 version of 02901 Advanced Topics in Machine Learning

For further information, please contact:

• Ph.D. Student Pola Schwöbel, E-mail: posc@dtu.dk
• Associate Professor, Ph.D. Morten Mørup, E-mail: mmor@dtu.dk, Phone: +45 45253900

DTU Compute, DTU, Matematiktorvet, Building 321, DK-2800 Kongens Lyngby Phone: +45 45253908, Fax: +45 45872599 January 2010 © 2010 Technical University of Denmark