Computer simulations are increasingly used in both biological research and therapeutic discovery, having significant potential to reduce animal usage, and inform drug development and clinical trial design. To fully exploit this potential it is critical that uncertainties in simulation design and experimentation are identified and quanitified. Current uncertainty analysis techniques are intractable for complex and resource intensive simulations, limiting their impact and application. Exploiting artificial intelligence techniques and well understood case studies, a novel computational framework will be developed that can significantly expedite and enrich statistical and uncertainty analyses to increase simulation confidence and adoption.
- Produced additional functionality for our spartan software tool, that is open source, available online from the R repository, and which attracts a high number of downloads.
- Found broader application in robotics, this interest now being exploited with the recent URF grant above.
- Been included as part of a workshop given at the University of Swansea, been presented at a conference at the University of Sheffield, and demonstrated at an invited talk at the University of Michigan.
- Investigating commercialisation of the developed approaches through potential licensiing agreements with external partners.
In addition, CFH outcomes used to inform a successful application for University Research Pump-Priming fund, that translates the expertise developed in analysing simulations of biological systems into the provision of platforms for gaining insights into the behaviours of simulated robotic systems.
New funding titled "Robo-Spartan: Machine Learning, Statistical Analysis, and High-Performance Computing Infrastructure for Designing and Understanding Robotic Systems." (not showing on PURE)