PhD scholarship in Distributed Optimisation with Applications to Federated Learning

This scholarship provides a stipend of $33,826 per annum (pro-rata) for three years and successful applicants will also be awarded a Tuition Fee Scholarship.

Applications are now open.

Until the position is filled.

One (1) scholarship available

The minimum requirements are:

• a bachelor's degree requiring at least four (4) years of full-time study in a relevant discipline awarded with honours. The degree should include a research component comprised of a thesis, other research projects or research methodology subjects that constitute at least 25% of a full time (or part time equivalent) academic year. The applicant must have achieved at least a distinction average in the final year; or 
• a master's degree that includes a research component comprised of at least 25% of a full time (or part time equivalent) academic year with an overall distinction average or a master's degree without a research component with at least a high distinction average; or 
• evidence of appropriate academic qualifications and/or experience that satisfies the Associate Deputy Vice-Chancellor Research Training and Development or nominee that the applicant has developed knowledge of the field of study or cognate field and the potential for research sufficient to undertake the proposed program.  At RMIT a grade of Distinction represents academic achievement of 70% or higher and a High Distinction is 80% or higher.

Interested candidates should contact the RMIT supervisor by email at minh.dao@rmit.edu.au. Please provide the following documentation to the RMIT supervisor in your email:

• a statement to outline your interest;
  
• an electronic copy of your academic transcripts;
• a CV that includes any publications and the contact details of 2 referees.

Upon response and approval from the supervisor, you should then submit an application for 'Admission Only' via the Online application service.

This PhD scholarship is funded by the Australian Research Council (ARC) Discovery Project "Distributed Optimisation without Central Coordination" whose objectives are to:

• Develop and analyse algorithms for distributed optimisation that do not require a central coordinator using monotone operator theory;

• Understand fundamental properties, subtleties and peculiarities of these methods for the purpose of informing distributed computation;

• Apply the newly developed algorithms to large-scale optimisation problems arising in federated learning.