This case study by Bangli Liu and Eufrásio Lima Neto (Module leader and lecturer, respectively, for Big Data Applications – CSIP5203) provides an overview of a seminar students engaged in on fairness and bias mitigation in ML models. Discussion was based on the paper “Identifying early help referrals for local authorities with machine learning and bias analysis”.
Article available at: https://link.springer.com/article/10.1007/s42001-023-00242-7
What happened?
Context: Discuss about important concepts about Legal, Privacy & Ethics in AI and Fairness and Bias Mitigation in ML algorithms
Description: A case study is presented addressing some topics about Legal, Privacy & Ethics in AI and Fairness and Bias Mitigation in ML algorithms
Evaluation: Students are involved in a discussion about the theme presenting your insights and thoughts.
Who was involved?
100 MSc Data Analytics students enrolled on the 7-week block module ‘Big Data Applications’ (CSIP5203).
What sustainability themes were covered?
Main theme: Social Justice and Ethics.
Lectures were delivered regarding Legal, Privacy & Ethics in AI and a seminar was held on Fairness and Bias Mitigation on ML algorithms.
Specific topics covered included:
- Legal Issues
- Privacy Issues
- Ethics Issues
- Fairness and Bias Mitigation
How has Problem Based Learning been used to embrace sustainability?
Part 1 – the following topics are covered:
- Legal concerns related to AI
- Deal with Legal concerns in AI
- Privacy Issues in AI
- Combat Privacy Issues in AI
- Ethics Issues
- Bias and Fairness
- Key Measures for Mitigating Bias
- Data Manipulation and Visualization
Part 2 – A case study is presented addressing some topics covered on Part 1
What are the successes from this work?
It’s important that students understand and reflect about the responsibility, consequences and impact of the AI and ML algorithms on the society.