Adaptive, Multi-Factor balanced, Regulatory Compliant Routing ADM Systems
Project details
This project develops new approaches to combine fairness, transparency and safety guarantees for ADM systems, such as machine learning based systems. We focus on resource allocation problems where there is a high level of uncertainty about the demand for resources, such as in the response to natural disasters or cyber security incidents.
In particular, we consider the problem of how criminal and malicious agents can manipulate such decision-making problems for their own advantage, and what measures can be taken to detect this manipulation.
University of Melbourne Researchers
- Professor Christopher Leckie (Lead Investigator)
- Dr Sarah Erfani (Associate Investigator