a*****g
发帖数: 19398 | 1 http://qz.com/723507/after-beating-a-human-at-go-the-deepmind-ais-next-challenge-is-preventing-blindness/
Diabetes can result in a common complication—known as diabetic retinopathy
—that damages the back of the eye, known as the retina. The chronic disease
can cause vision loss if left untreated. It’s one the leading causes of
blindness in American adults.
But, if caught early, patients can significantly reduce the risk of
blindness. Google’s DeepMind, which specializes in artificial intelligence,
is partnering with Britain’s National Health Service (NHS) to do just that
. The NHS is Britain’s publicly funded healthcare system—free to use,
treating a million patients every 36 hours, and consisting of one of the
largest workforces in the world.
“The NHS is a miracle,” says DeepMind’s co-founder Mustafa Suleyman. “
Not because it’s perfect, but because it works as well as it does within
some incredible constraints.”
Researchers will use machine learning to analyze anonymous eye scans, which
they will then use to create an algorithm that can better spot early signs
of eye conditions, such as diabetic retinopathy. The collaboration hopes to
develop eye scans that increase both the speed and accuracy of diagnosis.
Suleyman told The Guardian: “If you have diabetes, you’re 25 times more
likely to go blind. If we can detect this, and get in there as early as
possible, then 98% of the most severe visual loss might be prevented.”
Earlier this year, DeepMind sent shock waves after beating a human player at
the complex game of Go. The Alphabet-owned AI firm is keen to use the same
technology to revolutionize healthcare, but has sparked controversy in the
past.
This is the second collaboration between the NHS and the tech giant. The
first saw the development of a smartphone app that aimed to alert staff of
patients at risk of kidney failure. The data-sharing agreement between the
NHS and DeepMind was heavily criticized as Google was given access to the
healthcare data of up to 1.6 million patients.
Critics were concerned about the ways in which patient data would be shared
and whether it would remain secure. |
|
