Artificial Intelligence Breakthrough: ‘Better Than Human’ AI Algorithm Beats Brain In Test, An artificial intelligence algorithm designed to interpret images and video has beaten human scientists at mapping active neurons in the brain, a shocking new study has claimed.
Bio-engineers studying the complexities of the brain in mice chart out each individual neuron as it fires off during periods of activity. Using a process known as two-photon calcium imaging, scientists see bright spikes and flashes of light in the brain – nerve impulses firing off in response to stimulus. The process is incredibly slow, arduous and requires scientists to circle every single neuron, which lights up like a lightbulb. However, an incredible artificial intelligence algorithm tested by Duke University in North Carolina, US, has proven to be much more effective at the job.
A Duke study published this week in the Proceedings of the National Academy of Sciences has found the AI automated process is as accurate but considerably faster.
The algorithm, according to its lead author, appears to be “better than human experts”.
Duke University said in a statement: “This new technique, based on using artificial intelligence to interpret images, addresses a critical roadblock in neuron analysis, allowing researchers to rapidly gather and process neuronal signals for real-time behavioural studies.”
Typically, the human-led process of mapping neuron activity in a 30-minute-long video would take researchers anywhere from four to 24 hours.
Artificial intelligence trialled by Duke, on the other hand, was able to complete the process in mere minutes.
The AI proved faster and does not require food, hydration, toilet breaks or sleep to operate at peak performance.
Sina Farsiu, an associate professor of engineering at Duke, said: “As a critical step towards complete mapping of brain activity, we were tasked with the formidable challenge of developing a fast automated algorithm that as accurate as humans for segmenting a variety of active neurons imaged under different experimental settings.”
Fellow Duke researcher and professor Yiyang Gong said: “The data analysis bottleneck has existed in neuroscience for a long time – data analysts have spent hours and hours processing minutes of data, but this algorithm can process a 30-minute video in 20 to 30 minutes.
“We were also able to generalise its performance, so it can operate equally well if we need to segment neurons from another layer of the brain with different neuron size or densities.”
And Somayyeh Soltanian-Zadeh, a Duke PhD student and the study’s lead author said: “Our deep learning-based algorithm is fast and is demonstrated to be as accurate as, if not better than, human experts in segmenting and overlapping neurones from two-photon microscopy recordings.
Algorithm is fast and is demonstrated to be as accurate as, if not better than, human
Deep learning algorithms give scientists and researchers a chance to trawl through cast amounts of data with relative ease.
AI algorithms can be trained to identify different parts of a complex image for a specific purpose – in this case, to track firing neurons.
Duke’s researchers were so impressed with the algorithm’s ability to beat humans, they have made their software publicly available to the public.
READ MORE ON (Artificial Intelligence Breakthrough: ‘Better Than Human’ AI Algorithm Beats Brain In Test): EXPRESS