What it will take for AI to surpass human intelligence
Artificial Intelligence is already outsmarting us at '80s computer games by finding ways to beat games that developers didn't even know were there. Just wait until it figures out how to beat us in ways that matter.
Max Tegmark left his native Sweden in 1990 after receiving his B.Sc. in Physics from the Royal Institute of Technology (he’d earned a B.A. in Economics the previous year at the Stockholm School of Economics). His first academic venture beyond Scandinavia brought him to California, where he studied physics at the University of California, Berkeley, earning his M.A. in 1992, and Ph.D. in 1994.
After four years of west coast living, Tegmark returned to Europe and accepted an appointment as a research associate with the Max-Planck-Institut für Physik in Munich. In 1996 he headed back to the U.S. as a Hubble Fellow and member of the Institute for Advanced Study, Princeton. Tegmark remained in New Jersey for a few years until an opportunity arrived to experience the urban northeast with an Assistant Professorship at the University of Pennsylvania, where he received tenure in 2003.
He extended the east coast experiment and moved north of Philly to the shores of the Charles River (Cambridge-side), arriving at MIT in September 2004. He is married to Meia-Chita Tegmark and has two sons, Philip and Alexander.
Tegmark is an author on more than two hundred technical papers, and has featured in dozens of science documentaries. He has received numerous awards for his research, including a Packard Fellowship (2001-06), Cottrell Scholar Award (2002-07), and an NSF Career grant (2002-07), and is a Fellow of the American Physical Society. His work with the SDSS collaboration on galaxy clustering shared the first prize in Science magazine’s "Breakthrough of the Year: 2003."
Max Tegmark: I define intelligence simply as how good something is at accomplishing complex goals.
Human intelligence today is very different from machine intelligence today in multiple ways. First of all, machine intelligence in the past used to be just an always inferior to human intelligence.
Gradually machine intelligence got better than human intelligence in certain very, very narrow areas, like multiplying numbers fast like pocket calculators or remembering large amounts of data really fast.
What we’re seeing now is that machine intelligence is spreading out a little bit from those narrow peaks and getting a bit broader. We still have nothing that is as broad as human intelligence, where a human child can learn to get pretty good at almost any goal, but you have systems now, for example, that can learn to play a whole swath of different kinds of computer games or to learn to drive a car in pretty varied environments. And uh...
Where things are obviously going in AI is increased breadth, and the Holy Grail of AI research is to build a machine that is as broad as human intelligence, it can get good at anything. And once that’s happened it’s very likely it’s not only going to be as broad as humans but also better than humans at all the tasks, as opposed to just some right now.
I have to confess that I’m quite the computer nerd myself. I wrote some computer games back in high school and college, and more recently I’ve been doing a lot of deep learning research with my lab at MIT.
So something that really blew me away like “whoa” was when I first saw this Google DeepMind system that learned to play computer games from scratch.
You had this artificial simulated neural network, it didn’t know what a computer game was, it didn’t know what a computer was, it didn’t know what a screen was, you just fed in numbers that represented the different colors on the screen and told it that it could output different numbers corresponding to different key strokes, which also it didn’t know anything about, and then just kept feeding it the score, and all the software knew was to try to do randomly do stuff that would maximize that score.
I remember watching this on the screen once when Demis Hassabis, the CEO of Google DeepMind showed it, and seeing first how this thing really played total BS strategy and lost all the time.
It gradually got better and better, and then it got better than I was, and then after a while it figured out this crazy strategy in Breakout (where you’re supposed to bounce a ball off of a brick wall) where it would keep aiming for the upper left corner until it punched a hole through there and got the ball bouncing around in the back and just racked up crazy many points.
And I was like, “Whoa, that’s intelligent!” And the guys who programmed this didn’t even know about that strategy because they hadn’t played that game very much.
This is a simple example of how machine intelligence can surpass the intelligence of its creator, much in the same way as a human child can end up becoming more intelligent than its parents if educated well.
This is just tiny little computers, the sort of hardware you can have on your desktop. If you now imagine scaling up to the biggest computer facilities we have in the world and you give us a couple of more decades of algorithm development, I think is very plausible that we can make machines that cannot just learn to play computer games better than us, but can view life as a game and to do everything better than us.
Chances are, unless you happen to be in the Big Think office in Manhattan, that you're watching this on a computer or phone. Chances also are that the piece of machinery that you're looking at right now has the capability to outsmart you many times over in ways that you can barely comprehend. That's the beauty and the danger of AI — it's becoming smarter and smarter at a rate that we can't keep up with. Max Tegmark relays a great story about playing a game of Breakout with a computer (i.e. the game where you break bricks with a ball and bounce the ball off a paddle you move at the bottom of the screen). At first, the computer lost every game. But quickly it had figured out a way to bounce the ball off of a certain point in the screen to rack up a crazy amount of points. Change Breakout for, let's say, nuclear warheads or solving world hunger, and we've got a world changer on our hands. Or in the case of our computers and smartphones, in our hands. Max's latest book is Life 3.0: Being Human in the Age of Artificial Intelligence
Why do people with bigger hands have a better vocabulary? That's one question deep learning can't answer.
- Did you know that people with bigger hands have larger vocabularies?
- While that's actually true, it's not a causal relationship. This pattern exists because adults tend know more words than kids. It's a correlation, explains NYU professor Gary Marcus.
- Deep learning struggles with how to perceive causal relationships. If given the data on hand size and vocabulary size, a deep learning system might only be able to see the correlation, but wouldn't be able to answer the 'why?' of it.
One of the scientists with the Viking missions says yes.
- A former NASA consultant believe his experiments on the Viking 1 and 2 landers proved the existence of living microorganisms on Mars
- Because of other conflicting data, his experiments' results have been largely discarded.
- Though other subsequent evidence supports their findings, he says NASA has been frustratingly disinterested in following up.
Gilbert V. Levin is clearly aggravated with NASA, frustrated by the agency's apparent unwillingness to acknowledge what he considers a fact: That NASA has had dispositive proof of living microorganisms on Mars since 1976, and a great deal of additional evidence since then. Levin is no conspiracy theorist, either. He's an engineer, a respected inventor, founder of scientific-research company Spherix, and a participant in that 1976 NASA mission. He's written an opinion piece in Scientific American that asks why NASA won't follow up on what he believes they should already know.
Image source: NASA/JPL
Sunset at the Viking 1 site
As the developer of methods for rapidly detecting and identifying microorganisms, Levin took part in the Labeled Release (LR) experiment landed on Mars by NASA's Viking 1 and 2.
At both landing sites, the Vikings picked up samples of Mars soil, treating each with a drop of a dilute nutrient solution. This solution was tagged with radioactive carbon-14, and so if there were any microorganisms in the samples, they would metabolize it. This would lead to the production of radioactive carbon or radioactive methane. Sensors were positioned above the soil samples to detect the presence of either as signifiers of life.
At both landing sites, four positive indications of life were recorded, backed up by five controls. As a guarantee, the samples were then heated to 160°, hot enough to kill any living organisms in the soil, and then tested again. No further indicators of life were detected.
According to many, including Levin, had this test been performed on Earth, there would have been no doubt that life had been found. In fact, parallel control tests were performed on Earth on two samples known to be lifeless, one from the Moon and one from Iceland's volcanic Surtsey island, and no life was indicated.
However, on Mars, another experiment, a search for organic molecules, had been performed prior to the LR test and found nothing, leaving NASA in doubt regarding the results of the LR experiment, and concluding, according to Levin, that they'd found something imitating life, but not life itself. From there, notes Levin, "Inexplicably, over the 43 years since Viking, none of NASA's subsequent Mars landers has carried a life detection instrument to follow up on these exciting results."
Image source: NASA
A thin coating of water ice on the rocks and soil photographed by Viking 2
Levin presents in his opinion piece 17 discoveries by subsequent Mars landers that support the results of the LR experiment. Among these:
- Surface water sufficient to sustain microorganisms has been found on the red planet by Viking, Pathfinder, Phoenix and Curiosity.
- The excess of carbon-13 over carbon-12 in the Martian atmosphere indicates biological activity since organisms prefer ingesting carbon-12.
- Mars' CO2should long ago have been converted to CO by the sun's UV light, but CO2 is being regenerated, possibly by microorganisms as happens on Earth.
- Ghost-like moving lights, resembling Earth's will-O'-the-wisps produced by spontaneous ignition of methane, have been seen and recorded on the Martian surface.
- "No factor inimical to life has been found on Mars." This is a direct rebuttal of NASA's claim cited above.
Image source: NASA
A technician checks the soil sampler of a Viking lander.
By 1997, Levin was convinced that NASA was wrong and set out to publish followup research supporting his conclusion. It took nearly 20 years to find a venue, he believes due to his controversial certainty that the LR experiment did indeed find life on Mars.
Levin tells phys.org, "Since I first concluded that the LR had detected life (in 1997), major juried journals had refused our publications. I and my co-Experimenter, Dr. Patricia Ann Straat, then published mainly in the astrobiology section of the SPIE Proceedings, after presenting the papers at the annual SPIE conventions. Though these were invited papers, they were largely ignored by the bulk of astrobiologists in their publications." (Staat is the author of To Mars with Love, about her experience as co-experimenter with Levin for the LR experiments.)
Finally, he and Straat decided to craft a paper that answers every objection anyone ever had to their earlier versions, finally publishing it in Astrobiology's October 2016 issue. "You may not agree with the conclusion," he says, "but you cannot disparage the steps leading there. You can say only that the steps are insufficient. But, to us, that seems a tenuous defense, since no one would refute these results had they been obtained on Earth."
Nonetheless, NASA's seeming reluctance to address the LR experiment's finding remains an issue for Levin. He and Straat have petitioned NASA to send a new LR test to the red planets, but, alas, Levin reports that "NASA has already announced that its 2020 Mars lander will not contain a life-detection test."
Scientists discover the inner workings of an effect that will lead to a new generation of devices.
- Researchers discover a method of extracting previously unavailable information from superconductors.
- The study builds on a 19th-century discovery by physicist Edward Hall.
- The research promises to lead to a new generation of semiconductor materials and devices.