This application focuses on the development of computational systems for generating and evaluating health intervention strategies under uncertainty. The system is designed to explore structured intervention spaces, model trade-offs, and support decision-making where evidence is incomplete, heterogeneous, or context-dependent. The emphasis is on reasoning over options, constraints, and uncertainty rather than outcome prediction.
2025
We designed the decision logic and system architecture for representing symptoms, constraints, contextual factors, and intervention options as a unified problem space. The role involved formalising how different forms of evidence, historical use, and practical limitations interact in decision-making processes, and translating these interactions into computationally tractable structures.
We implemented multi-objective optimisation and evaluation frameworks to explore intervention spaces, assess trade-offs, and surface viable solution sets rather than single outputs. The system integrates structured knowledge representations, uncertainty-aware scoring, and constraint handling to support transparent, inspectable decision processes suitable for complex health domains.
Many health domains operate under conditions where definitive answers are unavailable, but decisions must still be made. This project demonstrates how decision-support systems can be designed to respect uncertainty, surface trade-offs, and make underlying assumptions explicit. While developed within a health context, the approach is applicable to other domains characterised by complex constraints, interacting factors, and incomplete evidence.
This application focuses on the development of computational systems for generating and evaluating health intervention strategies under uncertainty. The system is designed to explore structured intervention spaces, model trade-offs, and support decision-making where evidence is incomplete, heterogeneous, or context-dependent. The emphasis is on reasoning over options, constraints, and uncertainty rather than outcome prediction.
2025
We designed the decision logic and system architecture for representing symptoms, constraints, contextual factors, and intervention options as a unified problem space. The role involved formalising how different forms of evidence, historical use, and practical limitations interact in decision-making processes, and translating these interactions into computationally tractable structures.
We implemented multi-objective optimisation and evaluation frameworks to explore intervention spaces, assess trade-offs, and surface viable solution sets rather than single outputs. The system integrates structured knowledge representations, uncertainty-aware scoring, and constraint handling to support transparent, inspectable decision processes suitable for complex health domains.
Many health domains operate under conditions where definitive answers are unavailable, but decisions must still be made. This project demonstrates how decision-support systems can be designed to respect uncertainty, surface trade-offs, and make underlying assumptions explicit. While developed within a health context, the approach is applicable to other domains characterised by complex constraints, interacting factors, and incomplete evidence.
This application focuses on the development of computational systems for generating and evaluating health intervention strategies under uncertainty. The system is designed to explore structured intervention spaces, model trade-offs, and support decision-making where evidence is incomplete, heterogeneous, or context-dependent. The emphasis is on reasoning over options, constraints, and uncertainty rather than outcome prediction.
2025
We designed the decision logic and system architecture for representing symptoms, constraints, contextual factors, and intervention options as a unified problem space. The role involved formalising how different forms of evidence, historical use, and practical limitations interact in decision-making processes, and translating these interactions into computationally tractable structures.
We implemented multi-objective optimisation and evaluation frameworks to explore intervention spaces, assess trade-offs, and surface viable solution sets rather than single outputs. The system integrates structured knowledge representations, uncertainty-aware scoring, and constraint handling to support transparent, inspectable decision processes suitable for complex health domains.
Many health domains operate under conditions where definitive answers are unavailable, but decisions must still be made. This project demonstrates how decision-support systems can be designed to respect uncertainty, surface trade-offs, and make underlying assumptions explicit. While developed within a health context, the approach is applicable to other domains characterised by complex constraints, interacting factors, and incomplete evidence.