# Modelling Arguments :::info ℹ️ **About this Document** This document should be used for brainstorming and identifying themes related to modelling arguments and patterns, which have mutual interest for those involved. It follows an action point from a discussion betwen Steffen Zschaler (Kings College London), Carlos Gavidia-Calderon (Alan Turing Institute) and Christopher Burr (Alan Turing Institute). ::: ## Topics of Interest How to represent the relationship between evidence, claims, and goals, with specific emphasis on how to formalise or quantify the degree to which evidence supports claims and a set of claims assure a goal. **Topics**: - Bayesian Networks - Directed Acyclic Graphs - Formal Epistemology How to support both dialogical aspects of an assurance case (e.g. defeaters, contested claims), as well as representing the evolution of an argument as an iterative process embedded within a project lifecycle. How to maintain such assurance cases as explicit artefacts directly linked to simulations and other digital twin artefacts and how such explicit artefacts can be used to support digital-twin maintenance and evolution. **Topics**: - Explainability - Human-Centred Design How to evaluate LLMs in their role as augmenting the process of assurance. **Topics**: - Safety and Evaluation - Schema Development ## Resources - Finding the Pattern You Need: The Design Pattern Intent Ontology: https://link.springer.com/chapter/10.1007/978-3-540-75209-7_15 - Lakatos-style collaborative mathematics through dialectical, structured and abstract argumentation: https://www.sciencedirect.com/science/article/pii/S0004370217300267 ## People - Jason Konek (Department of Philosophy, University of Bristol) - Steffen Zschaler (Department of Informatics, Kings College London) - Carlos Gavidia-Calderon (Research Engineering Group, Alan Turing Institute) - Ibrahim Habli (Centre for Assuring Autonomy, University of York) - Christopher Burr (Turing Research and Innovation Cluster in Digital Twins, Alan Turing Institute) - Kalle Westerling (Turing Research and Innovation Cluster in Digital Twins, Alan Turing Institute) - Linda Sheard (Microsoft) - Geoff Keeling (Google DeepMind)