About the Project
About UQ
With unrivalled research expertise and an international reputation for excellence, The University of Queensland provides every advantage to succeed during your research studies.
Over 15,000 students have graduated from UQ with a higher degree by research. Join our dynamic research community and benefit from world class supervision, reputation and facilities.
About the Boeing - UQ Research Alliance
The Alliance involves 30 Boeing staff members working at UQ’s campus, and collaborating with UQ researchers and students on topics including: cabin disease transmission; unmanned aircraft and autonomous systems; and environmental monitoring technologies (including for the Great Barrier Reef).
UQ Vice-Chancellor and President Professor Peter Høj said it marked a global innovation leader’s vote of confidence in UQ. “This agreement is a strong endorsement from Boeing of UQ staff, students and graduates, following 13 years of collaborations which have enabled the company to test UQ’s capabilities in areas including neuroscience, maths and advanced engineering,” he said.
Boeing develops an aviation ecosystem simulator which can predict flight, aircraft, airport and weather attributes based on current and historical data. The simulator can run scenarios thousands of times faster than real-time and therefore generate multiple potential outcomes. Trust in the outcomes depends on identifying the range of naturally expected variability, and incorporating that variability effectively in the simulation. Currently, methods of variation use Monte Carlo simulation, which is a general purpose randomisation process. The scenarios generated are based on matching to historical data. Major computational and modelling challenges in this process include imputation of missing data, effective methods for matching to historical data, effective exploration of the range of real-world variation, and methods to evaluate levels of trust in the outcome probabilities. There are several areas where the simulator can be enhanced and evaluated by incorporating forms of real world variation into the modelling process.
About Project 1 - Machine learning methods for aviation optimisation
This project will use complex systems approaches to examine trust from the perspective of the observed multivariate distribution in simulation outputs. The process will begin with exploring the multivariate data distribution (including traffic movements and weather patterns). Marginal distributions of single variables then need to be joined with dependencies (either explicit or implicit) to produce a high quality representation. There is an inherent trade-off between multivariate dependencies, computational cost, the amount of data available, and confidence in the trained system. The project will evaluate the levels of trust that can safely be attributed to such systems using validation processes that compare different optimisation criteria.
This scholarship has a total annual value of $41,232, and is open to domestic and international applicants.
Applications close 22 July 2018.
Applicants with a degree in statistics, engineering, computer science, mathematics or physics, with relevant research experience are encouraged to apply.
Please visit this website for more information: https://graduate-school.uq.edu.au/boeing
Funding Notes
This scholarship has a total annual value of $41,232, and is open to domestic and international applicants.
Continue with Facebook
