Mind Your Behaviour: Computational Modelling of Emotion & Desire Regulation for Behaviour Change

The application of computational technologies in the regulation of humans’ mental and behavioural processes has increased the understanding of these processes. The insight gained into these processes through computational technologies provides new opportu-nities, for instance, using Artificial Intelligence (AI) for mental health and behavioural well-being. Likewise, computational models of such processes could provide a base for intelligent systems that can support the regulation of unhealthy behaviour and lead to a healthy lifestyle. Therefore, gaining knowledge of human mental processes is not only promising for the development of more sophisticated cognitive and or behaviour change computational models but also for the development and deployment of real-time intel-ligent support systems based on the models. For instance, prior to the development of a system for emotions and desire regulation, getting insight into its generation is as essential as its regulation. Emotions are considered a rapid information-processing sys-tem and form a significant part of one’s behaviour. If a person lacks regulatory capabili-ties in terms of negative emotions or desires, they will most probably end up in bad mental or physical health. Similarly, getting to understand the contextually adaptive use of emotion regulation strategies and their interaction with desire is also of utmost importance. A person lacking flexible and adaptive use of regulation strategies is as vulnerable to mental and physical health setbacks as a person lacking regulation capa-bilities. It can lead to but is not limited to problems like stress, depression, and or overweight and obesity. The development of such cognitive computational models, therefore, needs to take all the relevant literature into consideration for the accurate rep-resentation of these phenomena. Moreover, reflection of these models in an autonomous system needs a systematic procedure to ensure its accurate conversion. Therefore, in this thesis, in the first place, we explore computational models and analyse various aspects of emotions & desires and their regulation to further increase understanding of these cognitive processes. For instance, during the modelling and analysis phase, it focuses on the generation and regulation of emotions and desires and multi-order adaptation mechanisms in emotion regulation strategies over time. Besides, it also studies the interaction of emotions with (food) desires and the established ways to disentangle this interaction through Behaviour Change Interventions (BCIs) during the analysis phase. Moreover, after a thorough understanding obtained through computational modelling and analysis, this dissertation, in the second place, develops a theoretical framework to help translate these computational models into a standalone application for human sup-port in real life. Additionally, a prototype cognitive computational model-based behav-iour change system translated through the framework has been used for a user study. The model-based behaviour change system reports encouraging results in terms of food intake triggered by negative emotions. This, on the one hand, forms an illustration of the applicability of the framework to real-world problems, on the other hand, it pro-vides evidence to posit that these computational models can provide a base for human support systems if implemented as a system, e.g., autonomous or semi-autonomous agents.