Posts by Collection

cv

portfolio

publications

(Ir)rationality in AI: State of the Art, Research Challenges and Open Questions

Published in ArXiv pre-print, 2023

The concept of rationality is central to the field of artificial intelligence. Whether we are seeking to simulate human reasoning, or the goal is to achieve bounded optimality, we generally seek to make artificial agents as rational as possible. Despite the centrality of the concept within AI, there is no unified definition of what constitutes a rational agent. This article provides a survey of rationality and irrationality in artificial intelligence, and sets out the open questions in this area. The understanding of rationality in other fields has influenced its conception within artificial intelligence, in particular work in economics, philosophy and psychology. Focusing on the behaviour of artificial agents, we consider irrational behaviours that can prove to be optimal in certain scenarios. Some methods have been developed to deal with irrational agents, both in terms of identification and interaction, however work in this area remains limited. Methods that have up to now been developed for other purposes, namely adversarial scenarios, may be adapted to suit interactions with artificial agents. We further discuss the interplay between human and artificial agents, and the role that rationality plays within this interaction; many questions remain in this area, relating to potentially irrational behaviour of both humans and artificial agents.

Download here

(Ir)rationality and Cognitive Biases in Large Language Models

Published in Royal Society Open Science, 2024

Do large language models (LLMs) display rational reasoning? LLMs have been shown to contain human biases due to the data they have been trained on; whether this is reflected in rational reasoning remains less clear. In this paper, we answer this question by evaluating seven language models using tasks from the cognitive psychology literature. We find that, like humans, LLMs display irrationality in these tasks. However, the way this irrationality is displayed does not reflect that shown by humans. When incorrect answers are given by LLMs to these tasks, they are often incorrect in ways that differ from human-like biases. On top of this, the LLMs reveal an additional layer of irrationality in the significant inconsistency of the responses. Aside from the experimental results, this paper seeks to make a methodological contribution by showing how we can assess and compare different capabilities of these types of models, in this case with respect to rational reasoning.

Download here

Game-theoretic agent-based modelling of micro-level conflict: Evidence from the ISIS-Kurdish war

Published in PLOS ONE, 2024

This article delves into the dynamics of a dyadic political violence case study in Rojava, Northern Syria, focusing on the conflict between Kurdish rebels and ISIS from January 1, 2017, to December 31, 2019. We employ agent-based modelling and a formalisation of the conflict as an Iterated Prisoner’s Dilemma game. The study provides a nuanced understanding of conflict dynamics in a highly volatile region, focusing on microdynamics of an intense dyadic strategic interaction between two near-equally- powered actors. The choice of using a model based on the Iterated Prisoner’s Dilemma, though a classical approach, offers substantial insights due to its ability to model dyadic, equally-matched strategic interactions in conflict scenarios effectively. The investigation primarily reveals that shifts in territorial control are more critical than geographical or temporal factors in determining the conflict’s course. Further, the study observes that the conflict is characterised by periods of predominantly one-sided violence. This pattern underscores that the distribution of attacks, and target choices are a more telling indicator of the conflict nature than specific behavioural patterns of the actors involved. Such a conclusion aligns with the strategic implications of the underlying model, which emphasises the outcome of interactions based on differing aggression levels. This research not only sheds light on the conflict in Rojava but also reaffirms the relevance of this type of game-theoretical approach in contemporary conflict analysis.

Download here

talks

Approaches to Knowledge lecture

Published:

As part of the Approaches to Knowledge lecture series, I gave a talk describing my research. In particular, I highlighted the interdisciplinary aspeects of my current reseach, and emphasised the contributions from each discipline.

Jan Tinbergen European Peace Science Conference

Published:

Presented ongoing work ‘Spread of Protest Onset: The role of urban centers in national level protests’ as part of the 22nd Jan Tinbergen European Peace Science Conference.

Approaches to Knowledge lecture

Published:

As part of the Approaches to Knowledge lecture series, I gave a talk describing my research. In particular, I highlighted the interdisciplinary aspeects of my current reseach, and emphasised the contributions from each discipline.

Conflict & Change PhD workshop

Published:

Presented ongoing work ‘Spread of Protest Onset: The role of urban centers in national level protests’ as part of the 2024 Conflict & Change PhD workshop.

teaching

Dissertation supervision

Undergraduate dissertation, UCL, Computer Science, 2023

Co-supervised an undergraduate dissertation titled ‘A Quantitative Analysis of the Fridays for Future Movement.’