HPC-AI has now become a bundled term due to the rapid progress of deep learning architectures, which are empowered by running AI applications at scale on supercomputing infrastructures. NCI empower the research that needs computing, storage, and cloud resources through various merit-based allocation schemes. However, most PhD students are not eligible as independent lead CI to apply for the national scheme. To address the growing demand for computing and storage resources of the HPC-AI application, as well as support early career researchers, NCI offers 100KSU with appropriate storage, and AUD $10,000 per year to support PhD programs through the Australian HPC-AI Talent Program. The extra fund could provide opportunities for PhD students to participate in international domain expertise conferences or supercomputing conferences.
We are looking forward to supporting AI applications on diverse scientific problems such as solving differential equations, accelerating molecular dynamics, predicting 3-dimensional protein structures, controlling nuclear fusion, or weather forecasting with higher resolution and accuracy. Some HPC-AI applications encounter a range of workflow bottlenecks. The performance can be limited by compute bound, I/O bound, memory bound, or a combination of the above, pushing the need for HPCD into the exascale range. The unprecedented representational ability of deep neural network structures can only be brought to bear on the most complex and urgent global problems by model training with ultra-large datasets on heterogeneous exascale HPCD architectures. Hence we would like to identify those hurdles and provide support to address them for the next generation of young researchers.
We aim to support 10 PhD students in 2023. The general rule is that the fund and compute resources are to be used within the period of PhD study.
Should you have any questions, please send emails to training.nci@anu.edu.au.
Eligible candidates need to go through the application process for the scholarship program. The selection is merit-based and the judging committee makes the decision independently.
Date | Steps |
---|---|
20 Dec 2022 | Application open |
31 Jan 2023 | Application close |
1 – 15 Feb 2023 | Judge committee review and discussion |
16 Feb 2023 | Scholarship result announcement |
Please provide your resume (2 pages maximum), two reference letters and your application (2 pages maximum) to training.nci@anu.edu.au as PDF attachments. Please use the Application Template file below.
studentship_application_NAME_ORG.docx
For successful scholarship recipients, program reviews are conducted each quarter for the duration of the program and a final report is required at the end of the program.
Date | Stages |
---|---|
Feb 2023 | Onboarding with training |
Mar 2023 | Project plan initial review |
May 2023 | Quarter 1 review |
Aug 2023 | Quarter 2 review and mid-term showcase |
Nov 2023 | Quarter 3 review |
Jan 2024 | Final report |
Feb 2024 | End of program presentation and symposium |
More details to be released. Subscribe to the NCI newsletter to stay up to date.