FLERA - Federated Learning Ecosystem for Research in Australia

Accelerating translational research through Federated Learning
Merging applied artificial intelligence (AI) with health research, and human imaging to accelerate diagnosis, and facilitate precision management, of a range of human diseases.

Bringing together applied artificial intelligence (AI) and health research as a  transformative technology is at the core of FLERA. 

Our Research

In 2020, the MRFF-funded TRANSCEND (TRranslating AI Networks to Support Clinical Excellence in Neuro diseases) project was established to overcome bench-to-bedside roadblocks, by creating a seamless transfer of patient data between AI R&D and clinical practice.

The TRANSCEND eco-system provides a rich research and development (R&D) environment for clinical applications and broad expertise to advance applied AI research, building upon the team’s previous R&D work in the CRC-P project 'AI: new smarts for the medical imaging industry'.

FLERA represents the natural evolution of TRANSCEND: our goal is to be the partner of choice for supporting the accelerated development and adoption of AI solutions in health that rely on federated learning for healthcare.

FLERA provides more than 5 years of expertise in AI-medical imaging R&D, resulting from partnerships between academia, industry and health networks. 

Federated Learning (FL) is a framework that enables multiple entities or sites to collaborate in solving a machine learning problem, such as the creation of an algorithm for detection or segmentation of abnormalities in medical images. Each entity keeps their raw data local, exchanging only model weights and therefore preserving the privacy of, for example, health data. 

 ‘FLERA Box’ is a hybrid engineering solution that, simply put, transforms your existing infrastructure.  FLERA Box comprises several elements that handle data flow, storage, annotation and federated learning. 

Delivering optimal outcomes securely and economically

Data security and economic feasibility are the central pillars of FLERA, enabling safe, translational research with minimum resources.  In response to the complex decentralised model training environments, we have developed deep research expertise in noise labelling, domain adaptation/transfer and incremental learning to deliver enhanced FL outcomes for our partners.

Contact flera.admin@sydney.edu.au for more information.

The FLERA Team

With a track record of successful research-health provider collaboration, the FLERA team consists of internationally recognised experts across neurology, radiology, basic and applied neuroimaging, computer vision and artificial intelligence, clinical trials and healthcare delivery, e-data infrastructure, and governance.  

The team’s interdisciplinary and strategic collaboration with key industry partners such as NVIDIA, Sydney Neuroimaging Analysis Centre (SNAC), and an extensive network of clinical partners give the team a unique position and strong advantage in translating state-of-the-art scientific findings into real-world clinical impacts.

Team members

Dr. Chenyu (Tim) Wang: FLERA Program Lead and Neuroimaging Scientist, University of Sydney

Prof. Michael Barnett: Professor of Neurology, University of Sydney and Royal Prince Alfred Hospital

Prof. Fernando CalamanteDirector of Sydney Imaging Core Research Facility, University of Sydney

A/Prof. Weidong Cai : Director of Multimedia Lab, School of Computer Science, USYD

Dr. Ryan Sullivan: Senior Product Owner for Research Platforms, ICT & School of Biomedical Engineering, University of Sydney

Dr. Andy Shieh: Chief software architect, Sydney Neuroimaging Analysis Centre

Lynette Masters: Managing Radiologist, I-MED radiology, Brain and Mind Centre

Mr. Dongang Wang, AI and FL Engineering

Dr. Linda Ly, Data annotation and imaging analysis

Mrs. Anne de Lacoste, FLERA Product manager

Dr. Heidi Beadnall, MS Clinical Research

Dr. Mariano Cabezas

Dr. Dongnan Liu

Mr. Geng Zhan

Our Partners