, , ,

The Spooky Secret Behind Artificial Intelligence’s Incredible Power

Spookily powerful artificial intelligence (AI) systems may work so well because their structure exploits the fundamental laws of the universe, new research suggests.

The new findings may help answer a longstanding mystery about a class of artificial intelligence that employ a strategy called deep learning. These deep learning or deep neural network programs, as they’re called, are algorithms that have many layers in which lower-level calculations feed into higher ones. Deep neural networks often perform astonishingly well at solving problems as complex as beating the world’s best player of the strategy board game Go or classifying cat photos, yet know one fully understood why.

It turns out, one reason may be that they are tapping into the very special properties of the physical world, said Max Tegmark, a physicist at the Massachusetts Institute of Technology (MIT) and a co-author of the new research.

The laws of physics only present this “very special class of problems” — the problems that AI shines at solving, Tegmark told Live Science. “This tiny fraction of the problems that physics makes us care about and the tiny fraction of problems that neural networks can solve are more or less the same,” he said. [Super-Intelligent Machines: 7 Robotic Futures]

Last year, AI accomplished a task many people thought impossible: DeepMind, Google’s deep learning AI system, defeated the world’s best Go player after trouncing the European Go champion. The feat stunned the world because the number of potential Go moves exceeds the number of atoms in the universe, and past Go-playing robots performed only as well as a mediocre human player.

But even more astonishing than DeepMind’s utter rout of its opponents was how it accomplished the task.

“The big mystery behind neural networks is why they work so well,” said study co-author Henry Lin, a physicist at Harvard University. “Almost every problem we throw at them, they crack.”

For instance, DeepMind was not explicitly taught Go strategy and was not trained to recognize classic sequences of moves. Instead, it simply “watched” millions of games, and then played many, many more against itself and other players.

Like newborn babies, these deep-learning algorithms start out “clueless,” yet typically outperform other AI algorithms that are given some of the rules of the game in advance, Tegmark said.

Another long-held mystery is why these deep networks are so much better than so-called shallow ones, which contain as little as one layer, Tegmark said. Deep networks have a hierarchy and look a bit like connections between neurons in the brain, with lower-level data from many neurons feeding into another “higher” group of neurons, repeated over many layers. In a similar way, deep layers of these neural networks make some calculations, and then feed those results to a higher layer of the program, and so on, he said.

To understand why this process works, Tegmark and Lin decided to flip the question on its head.

“Suppose somebody gave you a key. Every lock you try, it seems to open. One might assume that the key has some magic properties. But another possibility is that all the locks are magical. In the case of neural nets, I suspect it’s a bit of both,” Lin said.

One possibility could be that the “real world” problems have special properties because the real world is very special, Tegmark said.

Take one of the biggest neural-network mysteries: These networks often take what seem to be computationally hairy problems, like the Go game, and somehow find solutions using far fewer calculations than expected.

It turns out that the math employed by neural networks is simplified thanks to a few special properties of the universe. The first is that the equations that govern many laws of physics, from quantum mechanics to gravity to special relativity, are essentially simple math problems, Tegmark said. The equations involve variables raised to a low power (for instance, 4 or less).  [The 11 Most Beautiful Equations]

What’s more, objects in the universe are governed by locality, meaning they are limited by the speed of light. Practically speaking, that means neighboring objects in the universe are more likely to influence each other than things that are far from each other, Tegmark said.

Many things in the universe also obey what’s called a normal or Gaussian distribution. This is the classic “bell curve” that governs everything from traits such as human height to the speed of gas molecules zooming around in the atmosphere.

Finally, symmetry is woven into the fabric of physics. Think of the veiny pattern on a leaf, or the two arms, eyes and ears of the average human. At the galactic scale, if one travels a light-year to the left or right, or waits a year, the laws of physics are the same, Tegmark said.

All of these special traits of the universe mean that the problems facing neural networks are actually special math problems that can be radically simplified.

“If you look at the class of data sets that we actually come across in nature, they’re way simpler than the sort of worst-case scenario you might imagine,” Tegmark said.

There are also problems that would be much tougher for neural networks to crack, including encryption schemes that secure information on the web; such schemes just look like random noise.

“If you feed that into a neural network, it’s going to fail just as badly as I am; it’s not going to find any patterns,” Tegmark said.

While the subatomic laws of nature are simple, the equations describing a bumblebee flight are incredibly complicated, while those governing gas molecules remain simple, Lin added. It’s not yet clear whether deep learning will perform just as well describing those complicated bumblebee flights as it will describing gas molecules, he said.

“The point is that some ’emergent’ laws of physics, like those governing an ideal gas, remain quite simple, whereas some become quite complicated. So there is a lot of additional work that needs to be done if one is going to answer in detail why deep learning works so well.” Lin said. “I think the paper raises a lot more questions than it answers!”

Live Science.

3 Comments

Leave a Reply
  1. Wiki Leaks and Clinton A Summation

    Copy and Share Widely!

    Clinton was running a Government within a Government ignoring the Rule of Law. Obama has made sure that the USA Separation of Powers does not exist any more. An Islamic Soros Clinton Criminal Enterprise. Copy and Share widely !

    As of Friday afternoon, WikiLeaks has released a total of 10,169 emails from the hacked account of John Podesta, Hillary Clinton’s campaign chairman.

    Some of the top stories include:

    1) Hillary’a campaign team held back emails between Hillary and Obama after gov’t subpoena: Clinton Campaign Chairman, John Podesta, wrote and email to Cheryll Mills (attorney) about holding back the emails between President Obama and Hillary Clinton after receiving a subpoena to reveal all.

    2) Hillary breaks FEC rules by meeting with Super PAC: A hacked email released by WikiLeaks exposes that Clinton’s team was in a panic over a pending story the NY Times was about to break, revealing Clinton had illegally attended meetings with the Priorities USA super PAC,.”

    According to FEC regulations, SuperPACS are supposed to be completely independent of a campaign. “These expenditures may not be made in concert or cooperation with, or at the request or suggestion of, a candidate, the candidate’s campaign or a political party,” states the regulations.

    3) Hillary hates “everyday Americans.” John Podesta sent an email on April 19, 2015, stating, “I know she has begun to hate everyday Americans” but I think you should use it once the first time she says I’m running for president because you and everyday Americans need a champion.”

    4) Clinton’s dream is OPEN BORDERS: A quote from a speech Hillary Clinton gave to a Brazilian bank on May 16, 2013 revealed she wants open borders. She stated: “My dream is a hemispheric common market, with open trade and open borders, some time in the future with energy that is as green and sustainable as we can get it, powering growth and opportunity for every person in the hemisphere.” [05162013 Remarks to Banco Itau.doc, p. 28]

    5) Clinton admits USA can’t possibly vet all refugees: Speaking in Chicago to the Jewish United Fund on October 2013, Clinton said: “And particularly with Syria which has everyone quite worried, Jordan because it’s on their border and they have hundreds of thousands of refugees and they can’t possibly vet all those refugees so they don’t know if, you know, jihadists are coming in along with legitimate refugees.”

    6) Clinton Admits to Having Two Sets Of Policies: In a speech to the National Multi-Housing Council on April 24, 2013, Clinton said, “But if everybody’s watching, you know, all of the back room discussions and the deals, you know, then people get a little nervous, to say the least. So, you need both a public and a private position.” She reportedly earned $225,000 for the speech.

    7) Funding ISIS: Hillary Clinton wrote an email to John Podesta on August 17, 2014, acknowledging that the governments of Qatar and Saudi Arabia are funding ISIS.

    “While this military/para-military operation is moving forward, we need to use our diplomatic and more traditional intelligence assets to bring pressure on the governments of Qatar and Saudi Arabia, which are providing clandestine financial and logistic support to ISIL and other radical Sunni groups in the region,” she wrote. Meanwhile, those same governments have funded both the Clinton Foundation and Hillary Clinton’s 2016 election campaign, with between $10 to $25 million going to the Foundation as of 2008, the Saudis bragged they had funded 20% of Clinton’s presidential campaign.

    A Qatar ambassador asked to meet with Bill Clinton “for five minutes” to present him a $1 million check “for his birthday.”

    8) Clinton & Obama Tricked Americans On Iran Deal: Leaked emails that including “talking points” on the Iran nuclear agreement in the summer of 2015, where they were trying to sell Americans on the deal was discovered to contain misleading information.

    9) “We are screwed!” over conflicts with Clinton Foundation: Bill Clinton’s advisor, Doug Band, sent an email to John Podesta and Cheryl Mills in November 2011 about the negative publicity surrounding the connections with his firm Teneo and the Clinton Foundation, and said, “I’m starting to worry that if this story gets out, we are screwed.” Band said Chelsea Clinton was “acting like a spoiled brat kid who has nothing else to do but create issues to justify what she’s doing.”

    10) Bill & Chelsea nearly drove Foundation executive to suicide: In another email exchange in December 2011, Doug Band told how he had managed to prevent Clinton Foundation COO Laura Graham from committing suicide, over work-related stress caused by Bill Clinton and Chelsea Clinton.

    11) Drivers Licenses To Illegal Aliens: Email exchanges between Clinton staff verify that Clinton wants to give drivers licenses to all “undocumented immigrants.”

    12) Pay To Play Exposed: Evidence was leaked suggesting that during the term Hillary Clinton was secretary of state, the Obama administration rewarded individuals who gave large donations to either the DNC or Obama’s “Organizing for Action” organization with ambassadorship positions. An email dated May 18, 2016 asked the DNC if they could set up a brief call “to go over our process for handling donations from donors who have given us pay to play letters.”

    13) Fake Trump Craigslist Ad Scheme: Already plotting how they could take Trump down with the “woman card”, the DNC discussed how they could place a FAKE Craigslist ad, making it appear that Trump was advertising for “hot” women.

    14) Clinton went anti-Muslim: In 2008, Hillary used race-driven talking points to discredit Obama in the general election. The campaign pushed the narrative that Obama was no good for America because his father was Muslim, and that he grew up around Muslims in a nation highly populated with Muslims.

    15) Donors expect special favors: One email offers an independent “governance review” by a law firm that specializes in nonprofits. The review was requested by Mrs. Clinton’s daughter Chelsea who had concerns about the Clinton Foundation and potential conflict-of-interests with its donors, and was sharply opposed by her father’s friends and donors. The review, conducted in December 2010, concluded many of the foundation’s donors had “an expectation of a quid pro quo benefits in return for gifts.”

    16) Mrs. Clinton’s spokesman urged her to lie to the public on whether she sent classified information: Bloomberg News reported Mrs. Clinton’s campaign spokesman Brian Fallon urged her to issue a blanket denial that she ever sent classified information through her private email server because anything less could open her up to charges she broke the law. “We should not think it is fine to find something that ‘should have been classified at the time,’ ” Mr. Fallon, the spokesman, wrote to other top campaign officials on Aug. 22, 2015, as they debated a statement Mrs. Clinton would make on her use of an email server in her home. “Our position is that no such material exists, else it could be said she mishandled classified info,” Mr. Fallon wrote.

Leave a Reply

Your email address will not be published. Required fields are marked *