Summer School: Fairness of Algorithms

2 December 2020, by Internetredaktion

At the interface of computer science, social sciences and law a new research field is currently establishing itself, which addresses the fairness of algorithms. Foci of research are, by which mechanisms and to what extent decisions made by algorithms can contain systematic biases – particularly in the context of machine learning and artificial intelligence.

The urgency of this highly interdisciplinary research field is underlined by the tremendous expansionof applications of artificial intelligence and algorithm-based decisions in all aspects of societal organization.

With this summer school we want to discuss some aspects of this new research field and attempt a dialog about these topics across disciplinary boundaries.

Call for Application

We ask doctoral students and post-docs to apply.

The summer school is an opportunity for all lawyers who work on legal issues which involve AI elements and assumes a deeper analysis of the respective field of law and want to work on algorithmic biases. Please, present your research topic and your motivation for your application (not more than 4.000 characters). CV and list of publication (if possible) are required. Please, send your application to claudia.schubert"AT"uni-hamburg.de.

For data scientists and computer scientists with a focus on machine learning / artificial intelligence this summer school is an opportunity to understand the challenges of identifying algorithmic biases and their societal impact. We require a statement indicating your competence level in ML/AI and a brief statement outlining your motivation for participating in this summer school (not more than 4.000 characters). CV and list of publication (if possible) are required.

Place: Otto-Bagge-Kolleg, Sehlendorf

Time: 6.-10.7.2021

Deadline for applications: 1.5.2021

Organisation Team: Prof. Dr. Marc-Thorsten Hütt (Jacobs University), Prof. Dr. Claudia Schubert (Universität Hamburg)

Please, note that the summer school will take place in person. If the pandemic situation does not allow this, we will postpone the event as we need a close cooperation with direct interaction.

Program

Part 1. Fairness of Algorithms – Introductory presentations

Data Scientists:
- Basic principles of machine learning and algorithmic decisions
- Challenges of identifying biases in algorithmic decisions
Lawyers:
- Presentations on fairness aspects by means of examples (selection by the participants with regard to their research interests, e.g. labour law, finance, medical law, administrative law)
- working out common legal requirements, existing regulatory strategy, specific legal requirements

Part 2. Working groups

Brainstorming about a definition of algorithmic fairness
Brainstorming about model scenarios
Creation of biased data
Brainstorming about automatic identification of biases
Brainstorming about regulatory interventions
Implementation of simple bias identifiers
Implementation and test of regulatory interventions

Part 3. Research and regulatory perspectives

Data Scientists:
- Presentations about recent research findings on algorithmic biases
- Presentations about recent efforts in computer science to address algorithmic biases
Lawyers:
- evolution of the legal standards, potential regulators and regulatory strategies
- legal approach and/or ethical approach
- legal/ethical reaction on uncertainty of technical evolution

Timetable

	Day 1	Day 2	Day 3	Day 4	Day 5
09-11	Arival (from 10 am)	Presenatations Team Hütt	Presenatations Team Hütt	Presentations Team Schubert	Consequences for lawyers/
11-13	Preparation Lunch (from 12 o'clock)	Presentations Team Schubert	Working Groups	Working Groups	for data scientists
13-14	Introduction
14-16	Presenatations Team Hütt	Working Groups	Working Groups	Group Presentations	Summary
16-18	Presentations Team Schubert	Working Groups	Working Groups	Group Presentations	Depature