21 Mar The new name in real AI, for healthcare
The worlds of Biology and Artificial Intelligence (AI) change quickly. Right now we are seeing an extraordinary rate of change where biological understanding and AI converge to address the problems of Healthcare in today’s society and there is a new name bursting on to the scene in healthcare AI:
Intelligent OMICS Limited, a company which brings together the advances in Biology and AI.
Intelligent OMICS Limited is the new identity for the university spinout formerly known as CompanDX. “People ask us about the new name” laughs Dr Simon Haworth, CEO. “Are you implying that other AI companies are ‘dumb’, they say? In actual fact that is exactly what we are saying: Intelligent OMICS uses genuine artificial intelligence to analyse healthcare big data and to spot patterns that human researchers are incapable of identifying. We use a neural networks ‘learning’ approach, applying our patented I3 algorithm developed in house, to drive our work in personalised medicine. The new name reflects the breadth of what we can now do and differentiates us from the increasing number of companies offering AI and pseudo-AI packages as a solution to all manner of healthcare issues but without real biological relevance.”
A recent study by MMC Ventures suggested that as many as 40% of ‘healthcare AI’ companies do not in fact include any Artificial Intelligence or Machine Learning (‘The State of AI, 2019: Divergence’). Many so-called AI companies are applying traditional regression analysis or similar, or mimicking established healthcare processes.
But why does this matter? The issue is that the ‘dumb’ AI approaches continue to be constrained by existing human knowledge and miss the opportunity to develop new drugs and diagnostics.
The Intelligent OMICS analysis applies AI to create new knowledge. When studying a particular disease, approximately 50% of the key biomarkers that the Intelligent OMICS system identifies have not previously been linked with that disease. These biomarkers are not just symptoms of a disease state but are the key drivers or determinants of the disease and represent crucial new knowledge of the disease under study. Novel biomarkers can be used as new drug targets and the company’s ability to distil results down to the minimum viable number of biomarkers without compromising sensitivity or specificity mean that the results can easily be incorporated into laboratory-based or hand-help point of care diagnostics.
The Intelligent OMICS R&D team, led by AI expert and personalised medicine veteran Professor Graham Ball, has recently achieved a major breakthrough in development of its I3 technology – I3
stands for Intuitive Informed Intelligence. The system learns by repeated, sequential modelling of data, fitting non-linear patterns. Following the research breakthrough, analysis that would previously have taken six or seven months to complete can now be completed in two to three weeks. The system learns sequentially from application of up to 50 million models to a given dataset per hour.
“In practice the R&D gains achieved by Professor Ball and his team now enable us to address the most critical question in medicine today: what don’t we know about each disease that would enable us to diagnose it more quickly and treat or prevent it more effectively?” comments Dr Haworth. “For a disease like Alzheimer’s the key question might be ‘Can we develop an early diagnostic from analysis of blood samples?’ whilst for chronic diseases such as COPD the question might be ‘Can we develop a drug to halt progression of the disease?’ or ‘Can we develop a companion diagnostic to identify which drug is suitable for each individual patient?”
The healthcare market recognizes that we must base the creation of new drugs and diagnostics on analysis of healthcare big data, but doesn’t yet know how. Intelligent OMICS provides that capability.