Keynote speakers
Marina De Vos (University of Bath)
From normative systems to business rules
In human society, norms, policies, and laws serve as mechanisms to describe, guide, and regulate expected behaviour. These rules outline the desired conduct of individuals and specify rewards or penalties for compliance or violation. Similarly, these concepts can be applied to socio-technical systems, encompassing both human and software agents. Within such normative systems, agents possess autonomy, enabling them to decide whether to adhere to norms or deviate from them.
This talk delves into the representation and computational reasoning of norms, policies, and laws, while ensuring sufficient clarity for human participants involved in the system. The focal point is InstAL, a domain-specific language (DSL) designed to capture deontic concepts and the effects of agents' actions. Execution is facilitated through answer set programming, a declarative logic programming language. Together, they provide a powerful approach for modelling, verifying, monitoring, and revising norms within socio-technical systems.
To illustrate the practical implications, this presentation presents a case study that explores the compliance of business processes with specific aspects of the General Data Protection Regulation (GDPR). To seamlessly integrate with established practices in business process modelling and semantic web technology, an ODRL layer was developed on top of InstAL, facilitating smooth integration with existing workflows.
This talk aims to offer valuable perspectives on the potential of combining InstAL, answer set programming, and the ODRL layer to effectively model, verify, monitor, and revise norms within socio-technical systems, as exemplified through the GDPR compliance case study.
BIO
Marina De Vos is a senior lecturer/associate professor in artificial intelligence and the director of training for the UKRI Centre for Doctoral Training in Accountable, Responsible, and Transparent AI at the University of Bath. With a strong background in automated human reasoning, Marina's research focuses on enabling improved access to specialist knowledge, the logical foundations of AI systems, explainable artificial intelligence methods, and modelling the behaviour of autonomous systems.
In her work on normative multi-agent systems, Marina combines her interests in the development of software tools and methods, drawing from a diverse range of domains including software verification, logic programming, legal reasoning, and AI explainability, to effectively model, verify and explain autonomous agents. Currently, Marina's exploration involves systems that possess the ability to autonomously evolve through external and internal stimuli.
Huimin Dong (Sun Yat-Sen University)
Resolving the Paradox of Free Choice Permission: A Semantic Approach
The concept of free choice permission is commonly understood as follows: if it is permitted to do α or β, then it is permitted to do α and it is permitted to do β. This differs from a permission that simply implies the absence of prohibition. However, when applying monotonic reasoning to this type of permission, a permission to do α logically leads to a permission to do both α and β. Problems arise when we introduce a prohibition on doing β, resulting in a paradox of free choice permission. Various proof theory solutions have been proposed to address the nonmonotonic aspects of this issue. In this talk, I will present a semantic approach aimed at resolving this problem. Following the tradition of dynamic logic, I will focus on the concept of open reading as the semantic core of free choice permission. Furthermore, I will discuss how the inclusion of normality can be incorporated into this framework to resolve the free choice paradox.
BIO
Huimin Dong, an Assistant Professor at Sun Yan-sen University's Department of Philosophy (Zhuhai), specializes in developing formal models for normative reasoning. Her interdisciplinary research covers logic, philosophy, ethics, law, and AI, with a particular focus on deontic logic, nonmonotonic reasoning, and logic-based methods for AI ethics and law.
Lou Goble (Willamette University)
Preemption and Plausible Oughts
I present a problem for deontic logic, a puzzle to invite further investigation. It is the problem of preemption: Imagine a case in which (a) there is something, X, that might occur but really should not, but also (b) there is something else, Y, that, if done, would preempt X, i.e., if Y were done, then X would not occur, while (c) if Y were not done, then X would happen. Moreover, in this case, neither X nor Y is determined; it is even possible (d) that Y not be done and X not happen, though, in light of (c), that is a remote possibility, even far-fetched. We may suppose too (e) it would be better for Y to be done and X not occur than for Y not to be done and X occur. And yet, although it is a remote possibility, (f) having Y not be done and X not occur would be better still than for Y to be done and X not occur. Given (a)--(f), and especially (a), (b), (c) and (e), the inference to (g) that Y should be done, seems clear, despite (d) and (f). The problem for deontic logic is to explain the validity of that inference in a plausible way. I present four approaches to that problem, all within the framework of branching time structures designed to accommodate the indeterminism inherent in the case. One of these approaches is familiar but it fails to account for the inference. The others are offered as more realistic alternatives that work better. Yet each has its drawbacks, and there is more work to be done. My primary purpose is to encourage that research.
BIO
Lou Goble studied philosophy and logic at Oberlin College (B.A.) and the University of Pittsburgh (M.A., Ph.D.), where he worked with Wilfrid Sellars, Alan Ross Anderson, and Nuel D. Belnap, Jr., amongst others. He then taught philosophy at the University of Wisconsin, Madison, and the University of North Carolina, Chapel Hill, before retiring for a while to the wilds of Oregon. He emerged from the shadows to teach at Willamette University, until his full retirement. The author of The Kalevide (a novel) and editor of The Blackwell Guide to Philosophical Logic and, with J. J. Ch. Meyer, Deontic Logic and Artificial Normative Systems (Proceedings of DEON 2006), he continues to write, though not usually for publication, on questions in philosophical logic and the philosophy of language, and more less academic matters.
John Horty (University of Maryland)
Knowledge representation for computational normative reasoning
I will talk about issues involved in designing a machine capable of acquiring, representing, and reasoning with information needed to guide everyday normative reasoning - the kind of reasoning that robotic assistants would have to engage in just to help us with simple tasks. After reviewing some current top-down, bottom-up, and hybrid approaches, I will define a new hybrid approach that generalizes ideas developed in the fields of AI and law and legal theory.
Joint work with Ilaria Canavotto
BIO
John Horty received his BA in Classics and Philosophy from Oberlin College and his PhD in Philosophy from the University of Pittsburgh; he is currently a Professor of Philosophy at the University of Maryland with affiliate appointments in Computer Science and the Institute for Advanced Computer Studies. His interests include logic, artificial intelligence, ethics, epistemology, philosophy of language, and philosophy of law. Horty is the author of four books as well as papers on a variety of topics in logic, philosophy, and computer science. His work has been supported by three fellowships from the National Endowment for Humanities and several grants from the National Science Foundation, by visiting fellowships at the Netherlands Institute for Advanced Studies and at the Center for Advanced Studies in Behavioral Sciences at Stanford University, and more recently, by a Humboldt Research Award.
Réka Markovich (Université du Luxembourg)
A formal theory of rights
Deontic logics traditionally concern impersonal obligations, however, to understand some basic concepts and structures of law, agents must be explicitly taken into account. For understanding rights, one actually needs to consider pairs of agents and formally map the variants of relations between them. In the talk, I give a comprehensive overview of the formal theory of rights I have been working on in the last few years. This work contributes to the tradition of the theory of normative positions, but exceeds that by defining formal characterizations using a multi-modal language, by sketching a theory of legal metaphysics, and also by offering resolution to difficulties rights theories often face.
BIO
Réka Markovich researches computational legal theory and studies its applications in Artificial Intelligence and legal reasoning. Her focus areas are legal knowledge representation, normative multi-agent systems, deontic logic, machine ethics, and XAI. Réka has an interdisciplinary background: she has degrees in law, in logic, and in communications, and a PhD in logic. Réka is currently an independent research scientist at the Department of Computer Science at the University of Luxembourg where she is the head of the newly established Computational Law and Machine Ethics (CLAiM) group in the Interdisciplinary Lab for Intelligent and Adaptive Systems. She represents Luxembourg on the board of the Benelux Association for AI and in 2021, she got elected to the international Steering Committee of the Foundation for Legal Knowledge Based Systems.