Using our heads to advance AI

Since Google DeepMind’s algorithm AlphaGo crushed Lee Se-dol, one of the world’s best living Go players, artificial intelligence (AI) has been a hot issue in Korea.

A number of science and technology seminars on AI were quickly held, and the government pledged 3.5 trillion won ($3 billion) in research and development in the field for the next five years. And the nationwide interest is necessary, as Korea holds just 5 percent of the number of AI-related patents as the United States.

But to develop automated machines and computer programs that can more closely mimic actual people, research on the human brain itself must also take place.

This requires money, but investment in various brain research projects and institutes stopped at 125 billion won last year.
Other countries are already well ahead.

World research on the human brain took off two to three years ago. U.S. President Barack Obama in 2013 launched the so-called BRAIN Initiative — Brain Research through Advancing Innovative Neurotechnologies — pledging $3 billion for the next 10 years with the aim of completely mapping the brain.

The European Union also embarked on a decade-long “Human Research Project” to advance the fields of neuroscience, computing, and brain simulation with the ultimate goal of developing an artificial neural network at a cost of 1 billion euros ($1.1 billion).
Korea has also been investing in neuroscience, but it’s not enough. We need a national agenda.

Since 2014, Japan has been engaged in the Brain Mapping by Integrated Neurotechnologies by Disease Studies, or Brain/MINDS, project that studies the differences between the brains of humans and marmosets. Its annual budget of 3 billion yen ($26,000) may not be as high as its counterparts in the United States and European Union, but the narrow focus has given them the edge in the field of neuromapping.
Korea, meanwhile, is behind the front-runner United States in neuroscience by 5.7 years.

Korea must immediately begin its own version of the initiative. It may never be able to catch up if it dithers further. We can start now or pay huge royalties to use the proprietary neurotechnologies of advanced countries in the future.

Scientists believe the national project must meet three requirements if it is to join the global race in brain mapping.

The project must first be inclusive of both health care and the broader industry, and it must at least be as expansive and large as Japan’s Brain/MINDS project.

Second, we should invest specifically to study the Korean brain, because neural regions differ dramatically by race and environment.
Lastly, the project must last at least 10 years. The U.S. project was supposed to last only a decade and was extended by another two years, including budgetary increases. Most of Korea’s research and development projects on the brain are designed to last for five years at most.

Robert Greenberg, head of Second Sight, a leading company in the field of prosthetics for the blind, said it took him 25 years to complete his technology. He added that he never imagined it would take him so long.

If research on the relatively accessible retina took that long, researching the brain will undoubtedly require patience, time and money if we hope to achieve advances of any significance.
Translation by the Korea JoongAng Daily.

JoongAng Ilbo, March 29, Page B8.

*The author is head of the Center for Neuroscience at the Brain Science Institute, Korea Institute of Science and Technology.

by Rhim Hye-whon