When it comes to massive computing power, artificial intelligence, deep learning and medicine, eyes have been set on the sights of seeing how vision impairment can be detected and improved.
Indeed, the future looks brighter, thanks to developments in AI that will help ophthalmologists detect and diagnose retina disease earlier.
The news comes via Rob Makin, the general manager and group director of Lenovo's Data Centre Group business unit, who starts off by noting that "visual impairment or blindness affects more than 453,000 Australians" but, according to the World Health Organisation, "nearly 80% of all VI [cases] can be prevented or cured".
It's here we learn from him that "AI may hold the key to significantly reducing the impact of retinal diseases by helping ophthalmologists detect disease more effectively".
"In-person dilated eye examinations remain the standard method for detecting retinal disease despite being shown to have a fairly low accuracy rate.
"In collaboration with Lenovo, the Barcelona Supercomputing Centre (BSC) is exploring how AI can improve the accuracy of the screening process and potentially detect a retinal disease sooner".
Overcoming the data barrier and AI neural networks
Here, Makin explains that "an AI neural network is a form of AI processing that allows self-learning based on past experience. This is modelled on the layers of neurons in a human brain and can develop conclusions based on complex and even potentially unrelated datasets".
"To train an AI neural network to detect small anomalies such as issues with the human eye would normally require a large amount of clean data. Unfortunately, this is not available so an alternative approach is required.
"To tackle this challenge, Lenovo and the BSC are using transfer learning to develop the visual analysis skills the AI neural network needs before applying that ability to identify visual degeneration."
"In short", said Makin, "we can train AI without actually having a large volume of clean data on the specific issue of VI. This not only reduces the data required, but the time to train the algorithm to identify issues within the data.
"This technology can then be packaged for delivery on a mass scale. A simple smartphone application can enable users to compete in an intuitive, gamified program while taking an active role in improving retinal disease screening.
"By participating, users teach the AI neural network how to accurately identify retinal anomalies in images, knowledge that can then be applied to help ophthalmologists in the screening process."
Bringing light to the world
"For most people living in Australia", Makin makes the observation that "vision is an integral part of their lives".
"Applying machine learning models to identify the various pathologies that cause VI in the first step in reducing the impact that VI will have on peoples’ lives.
"By augmenting the human experience of performing a retina examine, we can work towards eliminating preventable VI in Australia, and the world", Makin concludes, making great sense in helping us to further open our eyes on the benefits that AI and deep learning bring to the modern world.
Finally, a search online shows that the Barcelona Supercomputing Centre is a customer of Lenovo DCG, with the "MareNostrum 4" supercomputer nestled inside the BSC on the campus of the Polytechnic University of Catalonia.
This supercomputer boasts more than 3400 Lenovo ThinkSystem servers each performing over two trillion calculations per second, for a total of up to 11.1 petaflop capacity, giving MareNostrum 4 the power to probe deep secrets and foster innovation from genomics to geophysics, from fusion to the future.
So, with AI powered by massive computing power, the eyes have it, and clearly, it's looking like a welcome sight for sore eyes!
Here's Lenovo DCG and BSC's video on the MareNostrum 4: