TRIBINA "BUDUĆNOST VEŠTAČKE INTELIGENCIJE"

[zakk]

Centar za promociju nauke u saradnji sa Studentskim kulturnim centrom, u utorak, 18. juna, u 19 časova, u Velikoj sali SKC-a organizuje tribinu sa aktuelnom, provokativnom temom „Budućnost veštačke inteligencije“. Polazeći iz različitih počenih pozicija, učesnici tribine ”Budućnost veštačke inteligencije” će biti u prilici da ukažu na određene fenomene nastupajuće veštačke inteligencije u oblasti robotike, etike, filozofije i umetnosti.


Činjenica da se nauka i tehnologija dramatično uvlače u sve delove našeg svakodnevnog života, izaziva krajnje spontanu i intuitivnu potrebu ljudi za naučno-popularnom edukacijom. Neophodnost tehničke pismenosti više nije pitanje interesovanja i senzibiliteta već polako postaje pitanje opstanka. Naučna otkrića ljudskog tela ili dubine svemira polako brišu granicu između fantastičnog i mogućeg. Naučno-fantastične priče su odavno počele da se obistinjuju, a mi ih svesno ili nesvesno podrazumevamo.


Učesnici tribine su:


Prof. dr Milan M. Ćirković, transhumanista
Dr Srđan Stanković, robotičar
Jelena Bogavac, rediteljka
Miljana Milojević, filozofkinja

[Jake Chambers]

Transhumanista?
Hm... bilo bi lepo da ima još neki informatičar, neki fizičar, i tako..

[zakk]

Ćirković je fizičar po struci.

[zakk]

I zapravo je samo on dovoljan za bilo koju temu 

[Jake Chambers]

To me već plaši.

[Mica Milovanovic]

Imaš se rašta i plašiti...

[дејан]

трансхуманиста и филозофкиња 

[Josephine]

филозофкиња 

a reci i nama da se i mi smejemo:)

[mac]

Imaš sagovornike of kojih jedan pokušava da prevaziđe ono što nas čini ljudskim, s ciljem da postane nešto više od pukog čoveka, a drugi insistira na podeli koja pokazuje koliko smo ipak ograničeni nekim zakonima biologije. Ima tu neke ironije...

[Meho Krljic]

Istraživač sa Univerziteta u Torontu je napisao interesantan rad o tome šta su u ovom trenutku prepreke kod kreiranja veštačke inteligencije i, između ostalog, argumentuje da je Turingov test relativno nepouzdan alat u ovoj priči jer ga je suviše lako prevariti. PDF:

http://www.cs.toronto.edu/~hector/Papers/ijcai-13-paper.pdf

[Alexdelarge]

[Meho Krljic]

Još malo wireda danas:


Siri’s Inventors Are Building a Radical New AI That Does Anything You Ask

[Meho Krljic]

Robots face new test of creative abilities

A US professor is proposing a new way to test whether artificial intelligence (AI) is on a par with that of humans.
Currently scientists use the Turing test - named after computer scientist Alan Turing - which evaluates whether an AI can convince a judge that it is human in a conversation.
Prof Mark Riedl, from the Georgia Institute of Technology, is proposing a new test.
It would ask a machine to create a convincing poem, story or painting.
Dubbed Lovelace 2.0 it is an iteration of a previous Lovelace Test, proposed in 2001.
Named after one of the first computer programmers, the original test required an AI to create something that it would be incapable of explaining how it was created.
Lovelace 2.0 develops that idea.
"For the test, the artificial agent passes if it develops a creative artefact from a subset of artistic genres deemed to require human-level intelligence and the artefact meets certain creative constraints given by a human evaluator," explained Prof Riedl.
The artefact could be painting, poetry, architectural design or a fictional story.
"Creativity is not unique to human intelligence, but it is one of the hallmarks of human intelligence," said Prof Riedl.
Algorithms have already created stories and paintings although according to Prof Riedl "no existing story generation system can pass the Lovelace 2.0 test".
 Inspiring music Experts had mixed feelings about how good such a test would be.
Prof Alan Woodward, a computer expert from the University of Surrey thinks it could help make a key distinction.
"I think this new test shows that we all now recognise that humans are more than just very advanced machines, and that creativity is one of those features that separates us from computers - for now."
But David Wood, chairman of the London Futurists, is not convinced.
"It's a popular view that humans differ fundamentally from AIs because humans possess creativity whereas AIs only follow paths of strict rationality," he said.
"This is a comforting view, but I think it's wrong. There are already robots that manifest rudimentary emotional intelligence and computers can already write inspiring music."
The 65-year-old Turing test is successfully passed if a computer is mistaken for a human more than 30% of the time during a series of five-minute keyboard conversations.
Back in June a computer program called Eugene Goostman, which simulates a 13-year-old Ukrainian boy, was said to have passed the Turing test although some experts disputed the claims.

[Meho Krljic]

Experts pledge to rein in AI research

Scientists including Stephen Hawking and Elon Musk have signed a letter pledging to ensure artificial intelligence research benefits mankind.
The promise of AI to solve human problems had to be matched with safeguards on how it was used, it said.
The letter was drafted by the Future of Life Institute, which seeks to head off risks that could wipe out humanity.
The letter comes soon after Prof Hawking warned that AI could "supersede" humans.
 Rampant AI AI experts, robot makers, programmers, physicists and ethicists and many others have signed the open letter penned by the non-profit institute.
In it, the institute said there was now a "broad consensus" that AI research was making steady progress and because of this would have a growing impact on society.
Research into AI, using a variety of approaches, had brought about great progress on speech recognition, image analysis, driverless cars, translation and robot motion, it said.
Future AI systems had the potential to go further and perhaps realise such lofty ambitions as eradicating disease and poverty, it said.
However, it warned, research to reap the rewards of AI had to be matched with an equal care to avoid the harm it could do.
In the short term, this could mean research into the economic effects of AI to stop smart systems putting millions of people out of work.
In the long term, it would mean researchers ensure that as AI is given control of our infrastructure, restraints are in place to limit the damage that would result if the system broke down.
"Our AI systems must do what we want them to do," said the letter.
The dangers of a rampant AI answerable only to itself and not its human creators was spelled out in early December by Prof Hawking when he said AI had the potential to "spell the end of the human race."
Letting an artificially intelligent system guide its own development could be catastrophic, he warned in a BBC interview.
"It would take off on its own, and re-design itself at an ever increasing rate," he said.

[Meho Krljic]

S obzirom da se mnogi mudrac podigao da nas upozori kako proizvodnja napredne artificijelne inteligencije može da upropasti sve što smo gradili ovolike milenijume, ovo je zgodan pregled trenutnih projekata koji se bave AI-jem i stanja u kome im se razvoj nalazi:


What you wanted to know about AI

[ALEKSIJE D.]

Ima li negde definicija inteligncije? Ako može na serbskoslovesnokom, da razmumemo mi nepristorojni, prljavi, neobrazovani i ljenji. Pa da krenomo od toga.

[Meho Krljic]

Ima, hvala Bogu i wikipediji, čak i na ćirilici:

Интелигенција или интелект (лат. intellectus) је ментална особина која се састоји од више способности: учење из искуства, адаптирање на нове ситуације, схватање и разумијевање нових ситуација и коришћења стеченог знања у интеракцији са окружењем.

[ALEKSIJE D.]

Aha, kapiram. Sad, kako to primentiti na mašinu da ti donese pivo a da zna da mora biti hladno i određene marke koju voliš, bez da joj kažeš... U protivnom da shvati kako ima da je polupaš sekirčetom.

[Meho Krljic]

I da, kad vidi da si u donjoj trećini flaše, odmah krene po sledeće!!!!!!!!!!!

[ALEKSIJE D.]

E, to! Čemu bi, inače, služila veštačka inteligencija kad čovek ni svoju ne ume da upotrebi? Prosečnom čoveku dovoljno je: 1. da zna da zakrsti potpis kada prima kredit;  2. da ne treba da vrši nuždu u gaće i po dvorištu; 3. da prepozna lik omiljenog političara i zna koji je broj kada treba da zokruži listić.
Ništa više.
Može li to sve mašina? Može. Još i bolje.

[Meho Krljic]

Google just announced that its AI has dreams and here's what they look like

[Meho Krljic]

UCSD scientists build brain-inspired computer

Inspired by the human brain, UC San Diego scientists have constructed a new kind of computer that stores information and processes it in the same place. This prototype "memcomputer" solves a problem involving a large dataset more quickly than conventional computers, while using far less energy, the scientists say in a study.
The memcomputer prototype is a specialized proof of concept, but can be improved into a general-purpose computer, say researchers led by Massimiliano Di Ventra, a UCSD professor of physics. Such memcomputers could equal or surpass the potential of quantum computers, they say, but because they don't rely on exotic quantum effects are far more easily constructed.
Besides solving extremely complex problems involving huge amounts of data, memcomputers can potentially teach us more about how the brain operates, Di Ventra said. While the brain is often compared to a computer, the two are organized and operate much differently.
The study was published Friday in the journal Science Advances. Di Ventra was senior author; first author was Fabio Lorenzo Traversa, also of UCSD.
Conventional computers store data in one location designated for memory, and transfer it to processors located elsewhere to computer answers. But the human brain combines storage and processing in one place, treating these as one combined entity.
First proposed a few years ago, memcomputers likewise combine the storage and processing functions. Called a "collective state," this complex signal actually contains the problem solution, which in theory can be easily extracted. The prototype demonstrates this can be accomplished.
Memcomputers theoretically surpass conventional computers in their ability to handle certain very large datasets. That's because for conventional computers, the complexity of the problem increases exponentially, while the dataset grows in a linear fashion. But for memcomputers, the complexity also increases in a linear fashion.
The difference between exponentiality and linearity can be illustrated by examining the differences between the numbers 10 and 100, as expressed in the number of digits in each number and their actual values. The first has two digits and the second three, a linear increase. But the increase in actual numerical value from 10 to 100 is exponential.
Problems that involve exponential increases in complexity quickly become impractical for conventional computing.
Networks of memcomputing units act unlike a computer but like a brain in that they can function well even when some parts don’t work, Di Ventra said by email.
Neurons often fail to function properly at the individual level, but the network can tolerate many errors and still function. By contrast, just one error in a program or a component can cause a computer to crash.
Studying this fault-tolerant property could teach us more about how brains work, and how they break down, Di Ventra said.
“From memcomputing we can learn for instance the ability of the network of interconnected memprocessors in bypassing broken connections, namely how robust is such a network to damage of its units while still able to compute specific tasks,” Di Ventra said. “This could possibly translate in our understanding of the maximum amount of damage to neurons done by degenerative diseases, like Alzheimer's, before we lose specific functions.”
The study represents a significant advance in the field, said Yuriy V. Pershin, another researcher who has collaborated with Di Ventra and Traversa, but did not take part in this study.
"To the best of my knowledge, this is the first time people have used the collective state approach to solve exponentially hard problems," said Pershin, an associate professor in the department of physics and astronomy at the University of South Carolina.
"This work opens the way of solving complex problems in one step in the analog fashion, potentially much faster (for large problems) compared to the time it takes to find the solution of the same problem by conventional computer algorithms," Pershin said by email.
The prototype memcomputer is limited because it is analog, not digital, Di Ventra said, also by email. Analog computing is especially susceptible to interference from noise, which limits the ability to scale up the numbers of memprocessors in one computer.
"However, memcomputers can be made also digital (namely to process 0s and 1s like our present computers) therefore less susceptible to noise and hence they are scalable to a large number of units," Di Ventra said. "These digital memcomputers then hold great promise to complement present computers in those tasks in which they are not efficient, such as the combinatorial problems we have considered in our paper."
Alternatively, the paper stated, noise can be reduced by adding error-correcting codes.

[Meho Krljic]

Dartmouth Contests Showcase Computer-Generated Creativity

Can an algorithm pass for an author? Can a robot rock the house? A series of contests at Dartmouth College is about to find out.
 Dartmouth is seeking artificial intelligence algorithms that create "human-quality" short stories, sonnets and dance music sets that will be pitted against human-produced literature, poetry and music selections. The judges won't know which is which.
 The goal is to determine whether people can distinguish between the two, and whether they might even prefer the computer-generated creativity.
 "Historically, often when we have advances in artificial intelligence, people will always say, 'Well, a computer couldn't paint a sunset,' or 'a computer couldn't write a beautiful love sonnet,' but could they? That's the question," said Dan Rockmore, director of the Neukom Institute for Computational Science at Dartmouth.
 Rockmore, a mathematics and computer science professor, spun off the idea for the contests from his experience riding a stationary bike. He started thinking about how the music being played during his spin class helped him pedal at the right the pace, and he was surprised when the instructor told him he selected the songs without the help of computer software.
 "I left there thinking, 'I wonder if I could write a program that did that, or somebody could?'" he said. "Because that is a creative act — a good spin instructor is a total artist. It sort of opened my mind to thinking about whether a computer or algorithm could produce something that was indistinguishable from or even perhaps preferred over what the human does."
 The competitions are variations of the "Turing Test," named for British computer scientist Alan Turing, who in 1950 proposed an experiment to determine if a computer could have humanlike intelligence. The classic Turing test involves intelligent computer programs that can fool a person carrying on a conversation with it, and there have been many competitions over the years, said Manuela Veloso, professor of computer science and robotics at Carnegie Mellon University and past president of the Association for the Advancement of Artificial Intelligence.
 There also have been many efforts to create computer programs that can generate music or text, she said. The Associated Press, for example, recently began using computer programs to automatically generate business earnings reports and plans to do the same for college sports stories.
 What sets the Dartmouth contests apart is the evaluation by judges who will try to determine whether the work was generated by computers or humans and whether people prefer the computer-generated work, Veloso said. She said it will be interesting to see who does the judging — fooling a publisher with a computer-generated short story would be more significant than fooling the average reader, for example.
 The contests, dubbed DigiLit, PoetiX and Algorhythms, will run through the upcoming academic year, with prizes awarded in April. For both writing contests, if a computer-generated story or poem is scored as human by a majority of judges, the creators will win $5,000, and a prize of $3,000 will be awarded to the team that enters the best software. In the music contest, six finalists will compete against human disc jockeys during a dance party, selecting music from a list of 1,000 tracks that will be released just before the competition. The prize again will be $3,000.

  Both the human DJs and the computers will be hidden from view of the dance floor, Rockmore said.
 "Roughly the way we have it imagined is like a 'Wizard of Oz' scenario — there's a curtain, and there could be someone behind there or not," he said. "Then people will score it, human or machine."
 The contests aren't intended to be dismissive of imagination, Rockmore said. Rather, he views them as another way of investigating the neuroscience of creativity and human taste.
 "It's a challenge to produce a moving piece of literature, and I think in challenging yourself to try to do it, it makes you think hard about what moves you about a story, what goes into making a story," he said. "Thinking hard about those questions has a long tradition in the humanities, so it's just another way in which we can bring another lens to that, not in order to replace it but just to honestly consider it."
 

[Meho Krljic]

Heh heh
 
 
 Microsoft's Caption A.I. Is Sometimes Right, Sometimes Stupid

[Meho Krljic]

Budućnost veštačke inteligenceije je, reklo bi se, gledanje sitkoma. Ali, u naučne svrhe, naravno:


Scientists force computer to binge on TV shows and predict what humans will do

Researchers have taught a computer to do a better-than-expected job of predicting what characters on TV shows will do, just by forcing the machine to study 600 hours’ worth of YouTube videos.
The experiment could serve as a commentary on the state of research into artificial intelligence, or on the predictability of sitcom plots. It also calls to mind the scenes from countless science-fiction movies where the alien gets up to speed on human culture just by watching TV.
MIT’s Carl Vondrick and his colleagues are due to present the results of their experiment next week at the International Conference on Computer Vision and Pattern Recognition in Las Vegas.
The researchers developed predictive-vision software that uses machine learning to anticipate what actions should follow a given set of video frames. They grabbed thousands of videos showing humans greeting each other, and fed those videos into the algorithm.
To test how much the machine was learning about human behavior, the researchers presented the computer with single frames that showed meet-ups between characters on TV sitcoms it had never seen, including “The Big Bang Theory,” “Desperate Housewives” and “The Office.” Then they asked whether the characters would be hugging, kissing, shaking hands or exchanging high-fives one second afterward.
The computer’s success rate was 43 percent. That doesn’t match a human’s predictive ability (72 percent), but it’s way better than random (25 percent) as well as the researchers’ benchmark predictive-vision programs (30 to 36 percent).
“Still a long way to go,” the team says in its own YouTube video.
The point of the research is to create robots that do a better job of anticipating what humans will do.
“It could help a robot move more fluidly through your living space,” Vondrick told The Associated Press. “The robot won’t want to start pouring milk if it thinks you’re about to pull the glass away.”
That predictive capability would also come in handy for robot caregivers. For example, if it looked as if someone was about to fall down, the robots could anticipate that and intervene, Vondrick said.
And who knows? Maybe they’ll be better able to rescue falling robots as well.

[Meho Krljic]

Google's AI can now learn from its own memory independently

[Meho Krljic]

Google's AI created its own form of encryption 
 

     Researchers from the Google Brain deep learning project have already taught AI systems to make trippy works of art, but now they're moving on to something potentially darker: AI-generated, human-independent encryption. According to a new research paper, Googlers Martín Abadi and David G. Andersen have willingly allowed three test subjects -- neural networks named Alice, Bob and Eve -- to pass each other notes using an encryption method they created themselves.      As the New Scientist reports, Abadi and Andersen assigned each AI a task: Alice had to send a secret message that only Bob could read, while Eve would try to figure out how to eavesdrop and decode the message herself. The experiment started with a plain-text message that Alice converted into unreadable gibberish, which Bob could decode using cipher key. At first, Alice and Bob were apparently bad at hiding their secrets, but over the course of 15,000 attempts Alice worked out her own encryption strategy and Bob simultaneously figured out how to decrypt it. The message was only 16 bits long, with each bit being a 1 or a 0, so the fact that Eve was only able to guess half of the bits in the message means she was basically just flipping a coin or guessing at random.
Of course, the personification of these three neural networks oversimplifies things a little bit: Because of the way the machine learning works, even the researchers don't know what kind of encryption method Alice devised, so it won't be very useful in any practical applications. In the end, it's an interesting exercise, but we don't have to worry about the machines talking behind our backs just yet. With open-source deep learning tools like Microsoft's Cognitive Toolkit, it might be interesting to see this play out on an even larger scale. 

[Meho Krljic]

Google’s AlphaGo AI secretively won more than 50 straight games against the world’s top Go players

[Meho Krljic]

CMU AI Is Tough Poker Player

As the "Brains vs. Artificial Intelligence: Upping the Ante" poker competition nears its halfway point, Carnegie Mellon University's AI program, Libratus, is opening a lead over its human opponents — four of the world's best professional poker players.
One of the pros, Jimmy Chou, said he and his colleagues initially underestimated Libratus, but have come to regard it as one tough player.
"The bot gets better and better every day," Chou said. "It's like a tougher version of us."
Brains vs. AI, which began Jan. 11 at Rivers Casino in Pittsburgh, pits Chou and three other leading players — Dong Kim, Jason Les and Daniel McAulay — against Libratus in a 20-day contest in which they will play 120,000 hands of Heads-Up, No-Limit Texas Hold'em poker. All four pros specialize in this two-player, unlimited bid form of Texas Hold'em and are considered among the world's top players of the game.
While the pros are fighting for humanity's pride — and shares of a $200,000 prize purse — Carnegie Mellon researchers are hoping their computer program will establish a new benchmark for artificial intelligence by besting some of the world's most talented players.
Libratus was developed by Computer Science Professor Tuomas Sandholm and his Ph.D. student, Noam Brown. Libratus is being used in this contest to play poker, an imperfect information game that requires the AI to bluff and correctly interpret misleading information to win. Ultimately programs like Libratus also could be used to negotiate business deals, set military strategy or plan a course of medical treatment — all cases that involve complicated decisions based on imperfect information.
In the first Brains vs. AI contest in 2015, four leading pros amassed more chips than the AI, called Claudico. But in the latest contest, Libratus had amassed a lead of $459,154 in chips in the 49,240 hands played by the end of Day Nine.
"I'm feeling good," Sandholm said of Libratus' chances as the competition proceeds. "The algorithms are performing great. They're better at solving strategy ahead of time, better at driving strategy during play and better at improving strategy on the fly."
Chou said he and the other pros have shared notes and tips each day, looking for weaknesses they can each exploit.
"The first couple of days, we had high hopes," Chou said. "But every time we find a weakness, it learns from us and the weakness disappears the next day."
The change from day to day is not unexpected, Sandholm said. Each night after poker play ends, the Pittsburgh Supercomputing Center's Bridges computer performs computations to sharpen the AI's strategy. During the day's game play, Bridges computes end-game strategies for each hand.
"The people at the Pittsburgh Supercomputing Center have done a great job," Sandholm said, noting the staff has moved workloads around to enable the computer to be used in the competition. Since the beginning of the contest, the center has increased the number of Bridges' computer nodes assigned to the poker tournament.
Play begins at 11 a.m. each day and ends after 8 p.m. The public is welcome to observe game play, which is in Rivers' Poker Room.
Brains vs. AI is sponsored by GreatPoint Ventures, Avenue4Analytics, TNG Technology Consulting GmbH, the journal Artificial Intelligence, Intel and Optimized Markets Inc. The School of Computer Science has partnered with Rivers Casino, the Pittsburgh Supercomputing Center through a peer-reviewed XSEDE allocation, and Sandholm's Electronic Marketplaces Laboratory for this event.           

[Meho Krljic]

Još vesti iz poker-orbite:

AI Decisively Defeats Human Poker Players

Humanity has finally folded under the relentless pressure of an artificial intelligence named Libratus in a historic poker tournament loss. As poker pro Jason Les played his last hand and leaned back from the computer screen, he ventured a half-hearted joke about the anticlimactic ending and the lack of sparklers. Then he paused in a moment of reflection.
“120,000 hands of that,” Les said. “Jesus.”
Libratus lived up to its “balanced but forceful” Latin name by becoming the first AI to beat professional poker players at heads-up, no-limit Texas Hold'em.  The tournament was held at the Rivers Casino in Pittsburgh from 11–30 January. Developed by Carnegie Mellon University, the AI won the “Brains vs. Artificial Intelligence” tournament against four poker pros by US $1,766,250 in chips over 120,000 hands (games). Researchers can now say that the victory margin was large enough to count as a statistically significant win, meaning that they could be at least 99.98 percent sure that the AI victory was not due to chance.“The algorithms we used are not poker specific. They take as input the rules of the game and output strategy”—Tuomas Sandholm, Carnegie Mellon UniversityPrevious attempts to develop poker-playing AI that can exploit the mistakes of opponents—whether AI or human—have generally not been overly successful, says Tuomas Sandholm, a computer scientist at Carnegie Mellon University. Libratus instead focuses on improving its own play, which Sandholm describes as safer and more reliable compared to the riskier approach of trying to exploit opponent mistakes.
“We looked at fixing holes in our own strategy because it makes our own play safer and safer,” Sandholm says. “When you exploit opponents, you open yourself up to exploitation more and more.”
Even more important, the victory demonstrates how AI has likely surpassed the best humans at doing strategic reasoning in “imperfect information” games such as poker. The no-limit Texas Hold’em version of poker is a good example of an imperfect information game because players must deal with the uncertainty of two hidden cards and unrestricted bet sizes. An AI that performs well at no-limit Texas Hold’em could also potentially tackle real-world problems with similar levels of uncertainty.
“The algorithms we used are not poker specific,” Sandholm explains. “They take as input the rules of the game and output strategy.”
In other words, the Libratus algorithms can take the “rules” of any imperfect-information game or scenario and then come up with its own strategy. For example, the Carnegie Mellon team hopes its AI could design drugs to counter viruses that evolve resistance to certain treatments, or perform automated business negotiations. It could also power applications in cybersecurity, military robotic systems, or finance.“We took such a beating”—Dong Kim, poker proThe Libratus victory comes two years after a first “Brains vs. Artificial Intelligence” competition held at the Rivers Casino in Pittsburgh in April–May 2015. During that first competition, an earlier AI called Claudico fell short of victory when it challenged four human poker pros. That competition proved a statistical draw in part because it featured just 80,000 hands of poker, which is why the Carnegie Mellon researchers decided to bump up the number of hands to 120,000 in the second tournament.
The four human poker pros who participated in the recent tournament—Jason Les, Dong Kim, Daniel McAulay, and Jimmy Chou—spent many extra hours each day trying to puzzle out Libratus. They teamed up at the start of the tournament with a collective plan of each trying different ranges of bet sizes to probe for weaknesses in the Libratus AI’s strategy that they could exploit. During each night of the tournament, they gathered together back in their hotel rooms to analyze the day’s worth of plays and talk strategy.
The human strategy of playing weird bet sizes had its greatest success in the first week, even if the AI never lost its lead from the beginning. Libratus held a growing lead of $193,000 in chips by the third day, but the poker pros narrowed the AI’s lead by clawing back $42,201 in chips on the fourth day. After losing an additional $8,189 in chips to Libratus on the fifth day, the humans scored a sizable victory of $108,775 in chips on the sixth day and cut the AI’s lead to just $50,513.
But Libratus struck back by winning $180,816 in chips on the seventh day. After that, the “wheels were coming off the wagon” for the human poker pros, Sandholm says. They noticed that Libratus seemed to become especially unbeatable toward the last of the four betting rounds in each game, and so they tried betting big up front to force a result before the fourth round. They speculated on how much Libratus could change its strategy within each game. But victory only seemed to slip further away.
One of the players, Jimmy Chou, became convinced that Libratus had tailored its strategy to each individual player. Dong Kim, who performed the best among the four by only losing $85,649 in chips to Libratus, believed that the humans were playing slightly different versions of the AI each day.
After Kim finished playing on the final day, he helped answer some questions for online viewers watching the poker tournament through the live-streaming service Twitch. He congratulated the Carnegie Mellon researchers on a “decisive victory.” But when asked about what went well for the poker pros, he hesitated: “I think what went well was…shit. It’s hard to say. We took such a beating.”
In fact, Libratus played the same overall strategy against all the players based on three main components:

• First, the AI’s algorithms computed a strategy before the tournament by running for 15 million processor-core hours on a new supercomputer called Bridges.
• Second, the AI would perform “end-game solving” during each hand to precisely calculate how much it could afford to risk in the third and fourth betting rounds (the “turn” and “river” rounds in poker parlance). Sandholm credits the end-game solver algorithms as contributing the most to the AI victory. The poker pros noticed Libratus taking longer to compute during these rounds and realized that the AI was especially dangerous in the final rounds, but their “bet big early” counter strategy was ineffective.
• Third, Libratus ran background computations during each night of the tournament so that it could fix holes in its overall strategy. That meant Libratus was steadily improving its overall level of play and minimizing the ways that its human opponents could exploit its mistakes. It even prioritized fixes based on whether or not its human opponents had noticed and exploited those holes. By comparison, the human poker pros were able to consistently exploit strategic holes in the 2015 tournament against the predecessor AI called Claudico.

By the end of the tournament, the poker pros had long since been resigned to their fate. Daniel McAulay, the last poker pro to finish his hands for the day, turned to an offscreen spectator and joked, “How much do I have to pay you to play the last 50 hands? Uhhhh, this is so brutal.”
The Libratus victory translates into an astounding winning rate of 14.7 big blinds per 100 hands in poker parlance—and that’s a very impressive winning rate indeed considering the AI was playing four human poker pros. Prior to the start of the tournament, online betting sites had been giving odds of 4:1 with Libratus seen as the underdog. But Sandholm seemed confident enough in the AI’s tournament performance to state that “there is no human who can beat Libratus.”
Despite the historic victory over humans, AI still has a ways to go before it can claim to have perfectly solved heads-up, no-limit Texas Hold’em. That’s because the computational power required to solve the game is still far beyond even the most powerful supercomputers. The game has 10¹⁶⁰ possible plays at different stages—which may be more than the number of atoms in the universe. In 2015, a University of Alberta team demonstrated AI that provides a “weak” solution to a less complex version of poker with fixed bet sizes and a fixed number of betting rounds.“People are worried that my work here has killed poker: I hope it has done the exact opposite”—Tuomas Sandholm, Carnegie Mellon UniversityBut as the defeated poker pros drifted away from their computer stations one by one, gloomy viewer comments floated up on the live stream’s Twitch chat window. “Dude poker is dead!!!!!!!!!!!” wrote one Twitch user before adding “RIP poker.” Others seemed concerned about computer bots dominating future online poker games: “its tough to identify a bot from online poker rooms? ppl are terified [sic].”
There is some good news for anyone who enjoys playing—and winning—at poker. Libratus still required serious supercomputer hardware to perform its calculations and improve its play each night, said Noam Brown, a Ph.D. student in computer science at Carnegie Mellon University who worked with Sandholm on Libratus. Brown reassured the Twitch chat that invincible poker-playing bots probably would not be flooding online poker play anytime soon.
Another Twitch user asked if poker still counts as a skill-based game. The question seemed to reflect anxiety about the meaning of a game that millions of people enjoy playing and watching: What does it all mean if an AI can dominate potentially any human player? But Sandholm told the Twitch chat that he sees poker as “definitely a skill-based game, no question.”
“People are worried that my work here has killed poker: I hope it has done the exact opposite,” Sandholm said. “I think of poker and no-limit [Texas Hold’em] as a recreational intellectual endeavor in much the same way as composing a symphony or performing ballet or playing chess.”
As the final day of the tournament wound down, the Carnegie Mellon professor thanked the online viewers for watching and supporting the competition. And he took the time to answer a number of lingering questions about the new AI overlord of poker.
“Does Libratus call me daddy?” Sandholm read aloud a Twitch chat question. “No, it can’t speak.”
Editor’s Note: This story was updated to to reflect the fact that the statistical significance for the Libratus victory was 99.98 percent and not merely 99.7 percent.

[Meho Krljic]

First ever blueprint unveiled to construct a large scale quantum computer

[Meho Krljic]

The Dark Secret at the Heart of AI

No one really knows how the most advanced algorithms do what they do. That could be a problem.

[Labudan]

AI programs exhibit racial and gender biases, research reveals

Machine learning algorithms are picking up deeply ingrained race and gender prejudices concealed within the patterns of language use, scientists say

https://www.theguardian.com/technology/2017/apr/13/ai-programs-exhibit-racist-and-sexist-biases-research-reveals?CMP=share_btn_tw

[Meho Krljic]

Ja ovu vest videh jutros ali na kraju niswam mislio da treba da je prenosim jer je nekako jasno da AI naprosto sledi materijal koji joj je dat za učenje. Ako je većina vesti koje se vezuju za većinu populacije neutralna ili pozitivna a za manjine negativna, i to je materijal koji se daje za učenje, onda ne treba očekivati drugačije "rezonovanje" od strane AI.

[mac]

Možda je fora u tome što su ljudi mislili da su hranili AI tekstovima u kojima nema bias, pa su sad zaključili da ipak ima, ali ga oni nisu primetili.

[Meho Krljic]

Google's AlphaGo retires from competition

    To say that AlphaGo had a great run in the competitive Go scene would be an understatement: it has just defeated the world's number 1 Go player, Ke Jie, in a three-part match. Now that it has nothing left to prove, the AI is hanging up its boots and leaving the world of competitive Go behind. AlphaGo's developers from Google-owned DeepMind will now focus on creating advanced general algorithms to help scientists find elusive cures for diseases, conjure up a way to dramatically reduce energy consumption and invent new revolutionary materials.
         Before they leave Go behind completely, though, they plan to publish one more paper later this year to reveal how they tweaked the AI to prepare it for the matches against Ke Jie. They're also developing a tool that would show how AlphaGo would respond to a particular situation on the Go board with help from the world's number one player.
While you'll have to wait a while for those two, you'll soon be able to watch 50 games AlphaGo played against itself when it was training. If you'll recall, the researchers had it play games on its own after feeding it samples, so it could keep getting better and better. Go player Gu Li says those matches are "like nothing [he's] ever seen before" and that they're "how [he] imagine games from far in the future." DeepMind will upload those 50 games in batches, with the first 10 already available on its website.
   

[Meho Krljic]

Artificial intelligence can now predict suicide with remarkable accuracy

When someone commits suicide, their family and friends can be left with the heartbreaking and answerless question of what they could have done differently. Colin Walsh, data scientist at Vanderbilt University Medical Center, hopes his work in predicting suicide risk will give people the opportunity to ask “what can I do?” while there’s still a chance to intervene.
Walsh and his colleagues have created machine-learning algorithms that predict, with unnerving accuracy, the likelihood that a patient will attempt suicide. In trials, results have been 80-90% accurate when predicting whether someone will attempt suicide within the next two years, and 92% accurate in predicting whether someone will attempt suicide within the next week.
The prediction is based on data that’s widely available from all hospital admissions, including age, gender, zip codes, medications, and prior diagnoses. Walsh and his team gathered data on 5,167 patients from Vanderbilt University Medical Center that had been admitted with signs of self-harm or suicidal ideation. They read each of these cases to identify the 3,250 instances of suicide attempts.
This set of more than 5,000 cases was used to train the machine to identify those at risk of attempted suicide compared to those who committed self-harm but showed no evidence of suicidal intent. The researchers also built algorithms to predict attempted suicide among a group 12,695 randomly selected patients with no documented history of suicide attempts. It proved even more accurate at making suicide risk predictions within this large general population of patients admitted to the hospital.
Walsh’s paper, published in Clinical Psychological Science in April, is just the first stage of the work. He’s now working to establish whether his algorithm is effective with a completely different data set from another hospital. And, once confident that the model is sound, Walsh hopes to work with a larger team to establish a suitable method of intervening. He expects to have an intervention program in testing within the next two years. “I’d like to think it’ll be fairly quick, but fairly quick in health care tends to be in the order of months,” he adds.
Suicide is such an intensely personal act that it seems, from a human perspective, impossible to make such accurate predictions based on a crude set of data. Walsh says it’s natural for clinicians to ask how the predictions are made, but the algorithms are so complex that it’s impossible to pull out single risk factors. “It’s a combination of risk factors that gets us the answers,” he says.
That said, Walsh and his team were surprised to note that taking melatonin seemed to be a significant factor in calculating the risk. “I don’t think melatonin is causing people to have suicidal thinking. There’s no physiology that gets us there. But one thing that’s been really important to suicide risk is sleep disorders,” says Walsh. It’s possible that prescriptions for melatonin capture the risk of sleep disorders—though that’s currently a hypothesis that’s yet to be proved.
The research raises broader ethical questions about the role of computers in health care and how truly personal information could be used. “There’s always the risk of unintended consequences,” says Walsh. “We mean well and build a system to help people, but sometimes problems can result down the line.”
Researchers will also have to decide how much computer-based decisions will determine patient care. As a practicing primary care doctor, Walsh says it’s unnerving to recognize that he could effectively follow orders from a machine. “Is there a problem with the fact that I might get a prediction of high risk when that’s not part of my clinical picture?” he says. “Are you changing the way I have to deliver care because of something a computer’s telling me to do?”
For now, the machine-learning algorithms are based on data from hospital admissions. But Walsh recognizes that many people at risk of suicide do not spend time in hospital beforehand. “So much of our lives is spent outside of the health care setting. If we only rely on data that’s present in the health care setting to do this work, then we’re only going to get part of the way there,” he says.
And where else could researchers get data? The internet is one promising option. We spend so much time on Facebook and Twitter, says Walsh, that there may well be social media data that could be used to predict suicide risk. “But we need to do the work to show that’s actually true.”
Facebook announced earlier this year that it was using its own artificial intelligence to review posts for signs of self-harm. And the results are reportedly already more accurate than the reports Facebook gets from people flagged by their friends as at-risk.
Training machines to identify warning signs of suicide is far from straightforward. And, for predictions and interventions to be done successfully, Walsh believes it’s essential to destigmatize suicide. “We’re never going to help people if we’re not comfortable talking about it,” he says.
But, with suicide leading to 800,000 deaths worldwide every year, this is a public health issue that cannot be ignored. Given that most humans, including doctors, are pretty terrible at identifying suicide risk, machine learning could provide an important solution.

[Labudan]

Ne mož čoek se ubije da mu ne smetaju čike u mantilima.

Kacam reko da nam algoritmi rade o glavi niste vjerovali.

[ridiculus]

Sad nam još ostaje da počnu da predvidjaju ubistva.

[Labudan]

Filipe Diče, Srbija ti kliče!

[Meho Krljic]

Facebook shuts down AI system after it invents own language

[lilit]

hit
mislim, ljudi vide stvari koje nisu tu a ne vide one koje jesu.
nije AI izmislio jezik nego su ljudi napravili loš AI. i zato je isključen.

[Meho Krljic]

ARTIFICIAL INTELLIGENCE POLICY: A ROADMAP

[Meho Krljic]

Prajmer o mašinskom učenju i veštačkoj inteligenciji na Mediumu, napisan za običan svet:


Machine Learning for Humans

[Аксентије Новаковић]

'Whoever leads in AI will rule the world’: Putin to Russian children on Knowledge Day

'Whoever leads in AI will rule the world’: Putin to Russian children on Knowledge Day

 Vladimir Putin spoke with students about science in an open lesson on September 1, the start of the school year in Russia. He told them that “the future belongs to artificial intelligence,” and whoever masters it first will rule the world.

“Artificial intelligence is the future, not only for Russia, but for all humankind. It comes with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world,” Russian President Vladimir Putin said.

However, the president said he would not like to see anyone “monopolize” the field.

“If we become leaders in this area, we will share this know-how with entire world, the same way we share our nuclear technologies today,” he told students from across Russia via satellite link-up, speaking from the Yaroslavl region.

During the 45-minute open lesson (the standard academic hour in Russia), Putin also discussed space, medicine, and the capabilities of the human brain, pointing out the importance of cognitive science.

“The movement of the eyes can be used to operate various systems, and also there are possibilities to analyze human behavior in extreme situations, including in space,” Putin said, adding that he believes these studies provide unlimited opportunities.

The open lesson was attended by students and teachers from 16,000 schools, Rossiyskaya Gazeta reports. The total audience exceeded one million.

https://www.rt.com/news/401731-ai-rule-world-putin/

[Meho Krljic]

Artificial intelligence pioneer says we need to start over

[Meho Krljic]

AI: This Decade’s Worst Buzz Word

[Meho Krljic]

Budućnost veštačke inteligencije je... da pravi male veštačke inteligencije:


Google's AI Built Its Own AI That Outperforms Any Made by Humans

[Meho Krljic]

Ili to, ili pornografija, naravno:


AI-Assisted Fake Porn Is Here and We’re All Fucked

[Meho Krljic]

US Congress is trying to define what artificial intelligence actually means