AISEC seminar: Norms, Policy and Laws: Modelling, Verification and Monitoring

This week's seminar is by Marina De Vos'.
 
Marina De Vos', University of Bath, UK
Norms, Policy and Laws: Modelling, Verification and Monitoring
Wednesday, 2nd December, 11:55 GMT.
https://zoom.us/j/736667631
 

Abstract: Norms, policies and laws all focus on describing desired behaviour of actors, whether they are humans, software agents or processes. Actors (may) have the autonomy to decide to adhere to this behaviour or deviate from it. The actor’s actions can be checked against the desired behaviour, and compliance or violation potentially respectively rewarded or penalised. This talk explores the use of answer set programming, a declarative programming language, for the modelling, verification and compliance monitoring of these norms, policies, and laws. This is looked at from design time in terms of specification, debugging and combining models but also once the system is running in terms of monitoring, enforcement, and adaptation. Starting from normative multi-agent systems, where desired behaviour of agents can be described through a set of norms governed by so-called institutions, we look at case studies in law and business process modelling.

Bio: After completed her PhD in Computer Science and the Vrije Universiteit Brussel, Belgium, Marina De Vos joined the Department of Computer Science at the University of Bath, UK. She is a Senior Lecturer and Director of Training for UKRI Centre for Doctoral Training in Accountable, Responsible and Transparent Artificial Intelligence. Her research area is knowledge representation and reasoning, using answer set programming to model human/agent decision-making. Currently, her work focuses on the modelling, the explanation, and the verification of normative and policy-based reasoning in the areas of legal and socio-technical systems. In these systems participants, human and computational agents’ behaviour is guided by a set of norms/policies that describe expected behaviour. Non-compliance can be monitored and penalised while compliance is rewarded. Through a formal model, and corresponding implementation, the behaviour of the entire system can be proven and explained. Beyond normative modelling, she is interested in the software development for AI systems in general and logic-based systems more specifically and their use in wider society. 

Generally, all AISEC seminars are open for all interested colleagues, see http://laiv.uk/laiv-seminars/ 
There is a mailing list to join, if you wish to get seminar reminders and updates.