Read UQ's press release on Team MachineGenes' work and its entry into the AI XPRIZE semifinals here. Of the ten semifinalists, Team MachineGenes is the only one based outside of North America and Europe.
Team MachineGenes has reached the semifinal stage of the IBM Watson AI XPRIZE. The top 10 semifinalists will compete at TED HQ in New York in February 2020, with the hope of winning the US$3 million prize in the finals in April.
Watch Nigel's interview at the AI for Good Global Summit.
Team MachineGenes' work was showcased at the exhibition for the AI for Good Global Summit in Geneva, Switzerland. Special thanks to Advance Queensland and The University of Queensland for their sponsorship of Team MachineGenes.
Watch Nigel's pitch as one of the top-10 teams at the Australian finals for the Startup World Cup 2019. Full video available here.
Read a recent press release from UQ about the Diabetes Neuromathix work, particularly as demonstrated by Team MachineGenes in the IBM Watson AI XPRIZE.
As of January 2019, Australia and Singapore have granted patents on phi-TEA, with patents pending in the USA, Europe, Canada and Hong Kong. All commercial diabetes-related applications of these patents are permanently exclusively licensed to DNx worldwide, including the rights to sub-licence and independently prosecute or defend the patents.
Team MachineGenes is now listed as one of the top-10 teams worldwide in the running for the IBM Watson AI XPRIZE. Out of 147 approved teams (plus a wildcard) in 2017, only 30 teams remain after successive rounds of expert peer review. Team MachineGenes is the only team left in the world outside Europe and North America.Images courtesy of AI XPRIZE.
The patent offices of Australia and Singapore have granted Evolving Machine Intelligence patents for its evolutionary AI, all diabetes applications of which are permanently exclusively licensed worldwide to DNx.
DNx is now a member of NVIDIA's Inception Program, an accelerator program helping startups during critical stages of product development, prototyping, and deployment.
DNx as a member of Team MachineGenes, is now one of 147 teams, selected out of a total of 683 applicants, in the running for the IBM Watson AI XPRIZE after a peer review by world experts in AI.
Patent application PCT/AU2015/050362 ('A system and method for modelling system behaviour') enters national phase in USA, Canada, Europe, Hong Kong, Singapore and Australia.
DNx was a finalist in the Johnson & Johnson Quickfire competition held in conjunction with the Queensland Government.
The Australian Patent Office, acting as the international examiner, issued its preliminary finding that all seventy (70) claims for the patent application PCT/AU2015/050362 ('A system and method for modelling system behaviour') are globally novel and inventive and have industrial application (hence may be expected to give rise to patents in relevant jurisdictions throughout the world). This has strong implications for the international patent-protection of the Neuromathix Artificial Pancreas and its derivative technologies.
A related company to DNx, Evolving Machine Intelligence Pty Ltd, lodged an international patent application for its novel machine-intelligent platform, phi-Textured Evolutionary Algorithms ('phi-TEA'), titled 'A system and method for modelling system behaviour' (PCT/AU2015/050362), with priority from 30 June 2014. Under its permanent IP licence, DNx enjoys exclusive worldwide rights to this technology as applied to diabetes; this technology underpins the most recent versions of the Neuromathix Artificial Pancreas.
A peer-reviewed paper outlining the successful simulation study of 2011-12 was published in the Journal of Diabetes Science and Technology (N.J.C. Greenwood & J.E. Gunton, 2014, A Computational Proof of Concept of a Machine Intelligent Artificial Pancreas Using Lyapunov Stability and Differential Game Theory, Vol.8(4), 791-806).
A sister company, Evolving Machine Intelligence (EMI), lodged a provisional patent over a new proprietary machine-intelligence platform called phi-Textured Evolutionary Algorithms (phi-TEA), an enhancement to the Neuromathix algorithms also invented by Greenwood.
The US Patent Office has awarded NeuroTech Research Pty Ltd a US patent (No.8626522) for ‘Condition Analysis’, protecting the building-blocks for a method of human-interfaced or machine-intelligent control of complex nonlinear systems in the fields of medical therapies and biotechnology. These building-blocks are central to the Neuromathix algorithms. We also have a US continuation patent pending for further aspects of ‘Condition Analysis’. DNx enjoys an exclusive global right for commercial exploitation of these patents in the context of diabetes and its therapies.
As of late 2013, JDRF Australia, via its Clinical Research Network, awarded a further grant to the team to help fund a proof-of-concept clinical study demonstrating the Neuromathix artificial pancreas in human subjects with diabetes. Initially to be administered by the Garvan Institute, this study is now to be done at Westmead Hospital, Sydney.
Neurotech Research (NTR) has two sets of patents pending internationally. The first patent application has been granted in the USA, Australia and Singapore, with another Continuation patent pending in the USA. The second patent application has been granted in Australia and is pending in the USA.
Preparation for a collaborative clinical study between the Garvan Institute of Medical Research and Diabetes Neuromathix, ‘Machine-Intelligent Artificial Pancreas System’, part-funded by JDRF Australia under its Australian Type 1 Diabetes Clinical Research Network Pilot and Feasibility Grants began, following ethics approval.
The June 2013 edition of Australasian Science magazine is running a feature article on the simulation demonstration of the Neuromathix Artificial Pancreas:
A licence agreement has been signed between NeuroTech Research Pty Ltd and Diabetes Neuromathix Pty Ltd (DNx), awarding to DNx a permanent exclusive worldwide licence for all commercial diabetes applications of the Neuromathix algorithms.
Following the successful demonstration of the prototype Neuromathix artificial pancreas software, a diabetes-dedicated spin-off company, Diabetes Neuromathix Pty Ltd, has been formed as the commercialisation vehicle for this technology, to take it through clinical trials, regulatory approvals, to market and beyond.
In other news, the US Patent Office has advised (December 2012) NeuroTech Research Pty Ltd that a wholly-owned key patent application, ‘Condition Analysis’, will be awarded a US patent in 2013.
The collaborative simulation-based demonstration of the Neuromathix artificial pancreas, by Dr Greenwood and Associate Professor Jenny Gunton of Sydney’s Garvan Institute of Medical Research, has just concluded, with spectacular results: see
(For the background to this study, see the January 2012 press release below. Further reporting in the Australian and international media is also available online.)
We are now capital-raising to fund a clinical study in 2013, in preparation for FDA-compliant clinical trials in 2014.
One of Australia’s leading diabetes researchers, Dr. Jenny Gunton of Sydney’s Garvan Institute of Medical Research and Dr. Nigel Greenwood, a Brisbane mathematician, believe they have the edge in the international efforts to develop one of the big goals of diabetes research – the development of software for an artificial pancreas.
Such is their optimism that they have recently been awarded an US $110,000 research grant by JDRF to provide ‘seed’ funding for highly innovative research with potential significant impact on accelerating its mission.
The JDRF is dedicated to finding a cure for Type 1 diabetes and improving the quality of life for people with diabetes, funding critical research throughout the world.
An Honorary Senior Fellow at the School of Mathematics and Physics at the University of Queensland, Dr. Greenwood has been working since 2007 on one of the big issues for people with diabetes – how to decide the appropriate insulin dose for them as individuals.
Current insulin medication regimes are based on general population models of diabetes, modified by diabetics themselves through trial and error.
‘However, each diabetic’s body is different. There is no one size fits all. Too little insulin fails to bring blood sugar levels down far enough and cellular damage results. Too much insulin – and blood sugar levels drop too far, possibly inducing a coma’ Dr. Greenwood said.
‘Furthermore the human body is dynamic, with its behaviour changing across the day and as a result of external factors, such as exercise and food – which means that blood sugar levels and responses to insulin medication fluctuate.’
‘This makes it difficult for diabetics to manage their blood glucose levels.’
People with Type 1 diabetes maintain their blood glucose levels using a continuous glucose monitor (CGM) and taking periodic measurements each day of blood glucose using finger prick tests. The problem is to translate the data from the CGM and blood tests, correlated with past dosages of insulin being injected or infused, to work out the optimum, personalised insulin dosage regime. Dr. Greenwood has found a way around the complexity of all this by turning to what he knows best – mathematics.
He has developed a series of mathematical algorithms based on those used in missile-guidance and industrial robotics. In 2009, his start-up company, Neuromathix (NeuroTech Research Pty Ltd), received a $50,000 Queensland Government Medical Devices Financial Incentive Program grant to finish constructing diabetes-specific versions of the algorithms and convert them into software. In 2010, the company received a $30,000 Queensland Government Proof of Concept grant to demonstrate a prototype to deliver personalised medication regimes to simulated diabetics. Exploiting the recent revolution in desktop supercomputers and turning to cloud computing as both a distribution vehicle and the way to handle vast amounts of complex information, Dr. Greenwood said their research to date had been extremely successful. ‘But we need to perfect it and demonstrate that it has the ability to do the job. And this is where JDRF comes into the picture.’
CEO of JDRF Australia, Mike Wilson, said ‘JDRF are driving globally an initiative called the Artificial Pancreas Project, involving a consortium of top diabetes scientists, doctors and engineers. The idea is to integrate continuous glucose measurement with insulin pumps to automatically control glucose levels. This would be life-changing for many of the people struggling daily with managing Type 1 diabetes, including 122,000 Australian children and adults.’
To do this requires software that, in effect, has the ability to think for itself. ‘There is worldwide interest in developing software that can do the job, and we are very pleased that Australia and Australian researchers are playing a leadership role in making it possible.’
Dr. Greenwood said what gave them the edge was that their Neuromathix software was “paranoid” when analysing and acting upon diabetic data. ‘It doesn’t assume. It seeks to know what it doesn’t know, so has the capacity to learn, adapt and respond immediately to the unexpected. It is machine-intelligent software.’
Dr. Greenwood and Dr. Gunton are also working with Westmead Hospital in Sydney, using human data mined from medical histories to fine-tune their Neuromathix software. He said he hoped they could prove the worth of the Neuromathix software to the JDRF so they could move to the next stage – human trials.
The Queensland Government's Department of Employment, Economic Development and Innovation has awarded Neuromathix funding, to assist in the capability demonstration of the prototype Neuromathix artificial pancreas.
In late 2009, Neuromathix was awarded funding from Queensland Health's Office of Health and Medical Research for development and testing of an application of the Neuromathix algorithms to type-1 diabetes.