Sydney Neuroimaging Analysis Centre (SNAC) presented with XENON and NVIDIA at a recent conference: AI Hospital – How Hospitals are Leveraging AI, Machine Learning and Robotics in Healthcare on 25-November 2021. SNAC has been a long-term XENON customer and they have made impressive innovations in the application of AI to medical imaging which made for an interesting session.
Dr Wang from SNAC explored solutions for preserving privacy when using AI for medical imaging, using a Federated Learning model.
The session examined key issues of implementing machine learning models in a health care setting, including:
- Traditional machine learning models have limitations in privacy and security of the source data sets.
- Data Heterogeneity – Medical data is inherently diverse (type, dimensionality etc). This poses a challenge if data is not independent and identically distributed across participants.
- Traceability and accountability requires a high level of execution integrity and traceability – while ensuring privacy and security.
- Localisations of systems architecture at each institution requires AI infrastructure to be available at each site – with standardised data, labelling and training protocol.
This video covers these issues and covers SNAC’s journey from Conceptualisation to Commercialisation. Presented by:
- Dr Ettikan Karuppiah, Director/Technologist at NVIDIA
- Dr Chenyu (Tim) Wang, Director of Operations at Sydney Neuroimaging Analysis Centre (SNAC)
- Dr Werner Scholz, CTO and Head of R&D, XENON
This 30 minute video is both instructive and informative for medical institutes looking to implement AI, and deliver real benefits to patients.