“CSIRO’s Australian e-Health Research Centre (AEHRC) has led the way by becoming the first public sector organisation in the world to publish a machine learning-based health product on the AWS Marketplace”
Researchers are constantly advancing analysis approaches gaining sophisticated insights from complex data. However, the knowledge remains hidden in academic publications or obscure software repositories and does advance the analytics performed in business or clinical settings. Moreover, especially when machine learning software is applied in clinical settings, there are currently little guardrails ensuring the technology is set-up optimally or privacy/security requirements are satisfied, having prompted the FDA to call for feedback on how to regulate Machine Learning/AI as a medical device. All of this is exacerbated by the accelerating speed of innovation, which leaves little time for traditional software license agreements.
Operating in this field, the AEHRC is pioneering a new way of commercialising health solutions through the use of cloud-based digital Marketplaces. As covered by the AFR and highlighted by Iain Rouse, Amazon Web Services (AWS) Public Sector Country Manager, “CSIRO is the first public sector agency globally to offer a genomic analysis research product through AWS Marketplace”. Enabling software procurement “at the speed of cloud” and with passing AWS’s accredited security standards, AEHRC, can offer scientific solutions to commercial entities with the appropriate guardrails in place. The benefit of increased deployment speed, security and convenience enables AEHRC to commercialise peer-reviewed and open-access software solutions transparently.
AEHRC’s first software solution on the AWS Marketplace is VariantSpark, a machine learning library for ultra-high dimensional clinical data. VariantSpark is 90% faster than other machine learning implementations and requires 80% fewer samples than traditional analysis to detect statistically significant signal. The marketplace offering was released end of 2019 and has since been reviewed by a user as “powerful” and fast.
Read more on the transformational bioinformatics blog.
Year Completed: 2019