PhD Student in Artificial Intelligence: Causal Modelling and Multi-Agent Systems

Scope and purpose of the position

The position is part of the research project Formal Epistemology of Human-AI Interactions, funded by the Swiss National Science Foundation (SNSF). The project studies the conditions under which interactions between humans and artificial intelligence (AI) systems are epistemically reliable. The project aims to develop formal methods to represent, analyse and mathematically optimise human–AI interaction protocols. It takes into account factors such as how disagreements are handled, human oversight, interaction costs, and epistemic and governance constraints. The work includes developing causal-probabilistic models and evaluating them through multi-agent simulations in two application domains: neurological intensive care and cybersecurity. The project involves an interdisciplinary team with expertise in mathematics, logic, formal epistemology and artificial intelligence. It works with IDSIA research groups specialising in causal-probabilistic modelling and computational methods. It is also carried out in collaboration with Delft University of Technology and the University Hospital Zurich.

Duties and responsibilities

• Developing causal-probabilistic models to represent and analyse human–AI interaction protocols.
• Studying and implementing mathematical optimisation methods for protocols, taking into account performance, costs, and epistemic and governance constraints.
• Developing multi-agent simulation environments to evaluate protocols in the project’s two application domains (neurological intensive care and cybersecurity).
• Contributing to the project’s interdisciplinary activities and to the evaluation of results with domain experts.
• Writing scientific publications and presenting results at leading international conferences and journals in artificial intelligence, computer science and philosophy (e.g. NeurIPS, AAAI, FAccT, Synthese and Minds & Machines).

Requirements

Technical background and skills

• Master’s degree in computer science, artificial intelligence, mathematics, data science or a related field.
• Strong mathematical knowledge, particularly in probability and statistics, and knowledge of artificial intelligence and machine learning methods.
• Basic knowledge in at least one of: probabilistic modelling, causal inference, mathematical optimisation, or multi-agent systems.
• Good programming skills, preferably in Python, and experience implementing and evaluating computational models.
• Interest in the epistemological and ethical aspects of AI and in interdisciplinary research on human–AI interaction.
• Aptitude for original scientific research, ability to work both in a team and independently, scientific curiosity and strong motivation for research.
• Excellent command of English, spoken and written.
• Employment is conditional on admission to and enrolment in the PhD in Informatics at the Università della Svizzera italiana (USI) or the USI-SUPSI PhD programme in Applied Data Science and Artificial Intelligence

Desirable but not strictly required

• Advanced knowledge of, or project or research experience in, at least one of: causal and probabilistic modelling, multi-agent systems, mathematical optimisation, or simulation of intelligent systems.
• Prior research experience, documented through academic projects, a Master’s thesis or other scientific activities.
• Scientific publications or preprints in international conferences and journals, preferably in areas relevant to the project.
• Experience with libraries and frameworks for probabilistic modelling, causal inference, mathematical optimisation or multi-agent simulation (e.g. pgmpy, PyMC, CVXPY, Google ADK or equivalent).

What SUPSI Offers

• Fixed-term contract.
• Fully funded PhD.
• The opportunity to carry out original doctoral research with a significant degree of scientific autonomy, at the intersection of artificial intelligence, mathematics and formal epistemology.
• An international, interdisciplinary research environment, with scientific supervision by experts in AI, mathematical modelling and epistemology, and collaborations with Delft University of Technology and the University Hospital Zurich.
• Opportunities for international research mobility, participation in conferences and workshops, and publication in leading venues.
• Advanced training and professional development opportunities within the PhD programme.