Course overview
This course covers the fundamentals of algorithmic fairness, the most common biases in machine learning, and the importance of ethical and legal principles in AI design, training, and deployment. It explores the unique characteristics of social media data and the impact of AI on social good. The course discusses recent examples of algorithmic fairness in social media, AI fairness tools to mitigate discrimination and bias in ML models, and various methods and studies that address social biases in graph datasets and NLP methods. The goal of the course is to educate and prepare data scientists and engineers to ensure that their AI models are ethical, fair, and safe for use in human-in-the-loop applications.
This course has been commissioned as part of our open funding call for Responsible AI courses, with funding from Accenture and the Alan Turing Institute.
Who is this course for?
The course is designed for all researchers in data mining, artificial intelligence and social science. The audiences are assumed to have basic knowledge of probability, linear algebra and machine learning. Knowledge of python is required.
Learning outcomes
By the end of this courses, learners will be able to:
- Demonstrate a familiarity with the principles of ethical considerations for AI systems, including definitions of fairness
- Critically consider various sources of data biases in machine learning lifecycle
- Detect and assess biases in social media in both datasets and trained machine learning models
- Demonstrate sufficient knowledge on techniques that are commonly utilized to promote fairness in graph mining
