Why Artificial Intelligence Matters
Conversations around Artificial Intelligence vary from total confusion along with hints of fear (me) to hour-long debates around what AI actually is, the reality in its power to disrupt and why so many companies are jumping on the AI bandwagon. The latter occurred as I sat down for a lesson in AI from three Propel 2020 founders, using their technical powers for good in restoring care to healthcare, improving diagnostics and reducing the risk of life-changing injuries caused by heavy machinery.
Dr Declan Kelly, co-founder of Eolas Medical, AI-powered clinical decision support systems for healthcare.
Brendan Digney, co-founder of Machine Eye, intelligent safety and perception systems for plant and heavy machinery.
Dr Alan Kennedy, the founder of Pulse AI, transforming cardiovascular care with Artificial Intelligence.
The Definition of ‘Real’ AI
“AI is quite difficult to define because of its broad scope. Experts even argue over the definition of AI,” Declan, founder of Eolas Medical explains. “There are people who have been working on this for over 25 years, people who we’ve met who refer to it as ‘good old fashioned AI’ as in, the process that a machine will automate for a human being. Then you’ve got subsets of AI such as Machine Learning where a machine learns to do something that it’s not specifically programmed to do. Then there are even deeper neural networks which are what Alan (Pulse AI) considers real AI.”
To those of us without a PhD in Machine Learning, the term deep convolutional neural networks may not be one you encounter often. Alan (Pulse AI) clarifies, “They’re loosely based on how the animal brain works. That’s how they formed the idea, as a mathematical representation of how people think. That’s why everyone is scared about them but really, the stuff we do (in AI) can’t think for itself, it doesn’t have any intuition. It can’t decide to do something that we haven’t programmed it to do.”
Breaking things down further, Machine Eye co-founder, Brendan explains, “In AI, we always talk about data but it’s just about pattern recognition. If you throw enough historical data at it, you can start to learn from the patterns in that data. The more data you have, the more opportunity you have to see patterns. So it might be numerical data like with Pulse AI and patterns of a heartbeat. In my case with Machine Eye, it’s optical data. Even though it’s an image, it’s still a pattern and it’s looking for trends and recognising things within those patterns.”
Why AI?
For Machine Eye it’s a combination of timing and industry opportunity.
“AI has always been a buzzword up until now but we’re at a point where the technology is actually there to really see the benefits of it. The agriculture sector is probably the sector with the most potential to be disrupted here. The world isn’t making any more land but we have a much bigger population and more demands on the sector, so the fact that AI can be brought in to optimise that production is really going to transform things. It’s going to remove inaccuracies. It’s going to increase output monitoring, crop monitoring, livestock monitoring. Ireland has always jokingly been called the green country but it’s one of the best countries in the world for agriculture and to test new systems for agriculture. There’s huge potential for the whole industry to be transformed and become a cutting edge industry as a result.”
We’ve seen healthcare openly embrace AI and experience success in certain areas of the sector but with direct experience working in A&E, founder, Declan Kelly is ready to expand on the opportunity that AI brings to better the healthcare system.
“From a healthcare point of view, there’s been a lot of strides made with computer vision so any specialities that have that kind of attribute have seen brilliant progress like radiology, dermatology and ophthalmology. There are some companies in America that have FDA approval for AI-based algorithms that predict whether someone has diabetes and how badly that affects their eye. Recently in the news was mammography for breast cancer screening. They’ve shown AI algorithms can be better than a panel of experts in that field. Even better than that is the combination of experts and AI. Now we’re trying to achieve the same sort of success but with decisions for non-visual based things. For example, how someone describes pain in their chest when they’re having a heart attack or when someone is examined by a doctor. We’re trying to make sense of that data.”
In terms of why AI matters within healthcare, Alan believes it comes back to efficiency and taking tasks away from doctors that are mundane and repetitive.
“Initially, I think that’s how it’s going to be adopted, to help them to do less of that and focus on important elements of being a doctor. In Pulsa AI, we focus on cardiac diagnosis and electrocardiogram processing. Traditionally, doctors will look at an electrocardiogram and make a diagnosis based on things they see but what we are trying to do is bring through a new generation of ECG interpretation programmes, using AI-based models for automated interpretations with a view to augment them and make them quicker and more efficient. The tagline is to improve access to efficiency and accuracy in cardiac diagnosis with AI. Success for us is about improving patient safety, improving patient care and improving cardiac diagnosis.”
The Route to Disruption
Eolas Medical focus on ensuring that the best decision is made for every patient, every time. “The emergency department is under massive pressure at the minute. One of the issues with this is that there is an inefficiency in decision making and we’re trying to solve that. It allows doctors and nurses to focus more on actually caring for the patient while the more mundane stuff is being taken care of and being automated which will then improve patient safety. There’s a doctor in the States, Eric Topol. He wrote the Topol report which advises the NHS of how they should digitise the NHS moving forward over the next 20 years. He talks about how AI will restore the care in healthcare. It allows you to actually care for the patient on a more meaningful level. If you have a 20-minute consultation, 15 minutes of that may be taken up by you being on the computer or updating records. AI has the potential to turn that 20-minute consultation into 18 minutes with the patient and then it runs in the background. It’s not replacing the doctor, it’s augmenting the doctor. That will restore care in healthcare to allow the doctor-patient relationship to grow again.”
Tackling issues within agriculture, Machine Eye reduces the rate of life-changing injuries and fatalities that happen as a result of heavy machinery. “The big thing about a farm or a site is that it (the solution) needs to be effectively invisible. If it interferes with productivity, how somebody does their work or with the machine itself it will not be accepted by the industry. AI allows us to circumvent that. I always analogise Machine Eye to an airbag in a car. You buy a car, you understand there’s an airbag fitted but you never see it work or you hopefully should never see it work. It’s really the same thing with Machine Eye. Instead of putting these big guards in place, instead of making it awkward for people to do their job, an airbag keeps you safe but it doesn’t make it awkward to drive the car. With Machine Eye it’s the same, we sit there quietly, we monitor the environment, we understand what’s happening and we only actually become potent whenever we need to which is a completely different approach to how safety and perception have been done in these industries up until now. It has the potential to become a cooperative system with people as opposed to a physical barrier or a physical stop and it is underpinned by the power that we can get from those algorithms.”
There is a growing belief that, after decades of promise, AI is finally becoming a mainstream reality. If this trend continues, advancement in the next ten years could exponentially transform one industry after another. Personally, I’m glad that the future is being created by entrepreneurs such as the three I met. It is only with a careful combination of human and machine that we will truly be able to harness the power that AI offers.
🗞 Keep up to date with Ignite NI news, events and programmes by signing up to our monthly newsletter: bit.ly/IgniteNINews.
👀 Follow us on @IgniteAccelNI to see what the Propel 2020 (#Propel2020) teams get up to + follow the teams’ personal pages.
