Apple c-ofounder and Dancing Queen, Steve Wozniak told Michigan State University graduates that the computers he’s helped create will soon make us irrelevant as a species.
Woz was given an honorary doctorate of engineering from MSU for his many accomplishments in the personal computing industry.
But he warned that every time a computer was used to do something, it was making humanity less relevant.
According to Digital Trends, he claimed that the cyborgs are winning and also made a bit of a Freudian slip by added that the “androids are winning!”.
Woz said that artificial intelligence will be more like a real person. Computers will be like a friend you would want to go and meet. Somebody that can talk to your face; somebody that can understand things and knows the kind of jokes you like. AI will be like someone that knows what sort of things to say.
One of the hurdles for AI are the human senses. A computer cannot create art, if it can’t sense things that a human understands, like the wind on a beach.
Computers can hear, see, touch and sense motion, just like our inner ear, he said
Woz added that pretty soon there will be holograms, which will be much better than 3D television.
“We’ve created a new species; no question. We’re creators and, like I said, we’re making ourselves less relevant,” Woz said.
All that is left for humanity is Woz’s equation for happiness: H = 3F. It stands for “Happiness equals food, friends, and fun.” He said humanity must try to have fun while it is still relevant.
Top boffin Leslie Valiant, of Harvard University, has been named the winner of the 2010 Turing Award for his efforts to develop computational learning theory.
Valiant won the “Nobel Prize for computing” and $250,000 for a lifetime of work bringing together machine learning and computational complexity.
His work has lead to breakthroughs in artificial intelligence as well as computing practices such as natural language processing, handwriting recognition, and computer vision.
Valiant’s work created several new subfields of theoretical computer science, and developed models for parallel computing.
More famously he was behind many of the mathematical foundations of computer learning, an area of study that has led to breakthroughs such as IBM Corp.’s Watson, the machine built to play “Jeopardy!”
In 1984 he published a paper called the “Theory of the Learnable,” in Communications of the ACM. It turned out to be a best seller as it put machine learning on a sound mathematical footing.
It also created a new area for research area known as Computational Learning Theory. In the article Valiant worked out a general framework and a few concrete computational models.
His method of “Probably Approximately Correct” (PAC) learning has become a standard model for studying the learning process and has been adopted by the sub-editors of TechEye.
Building on Valient’s work, other boffins have developed algorithms that adapt their behaviour in response to feedback from the environment.
What are they smoking over at IBM’s R&D labs? First it decides to enter a supercomputer on Jeopardy!, and now it has filed a patent for toys with advanced Artificial Intelligence (AI).
The company has put forward a patent by the name of: “Adaptive System for Real-Time Behavioural Coaching and Command Intermediation”. It wants to use the technology, and eventually the toys, to replace the need for boring Supernanny-esque reality shows.
In the patent it explains the invention could: “include multiple interaction operations that can be performed by the interactive device for helping the child play less rough with other children.”
Or, “For example, one interaction operation can include an audible warning telling the child ‘to play nice’ in a strict tone of voice, whereas another interaction operation can include an audible warning that asks the child ‘would you like someone to do that to you’ in a softer tone of voice along with a visual cue as well.”
The stems from the fact that the “emotional centre” for humans is the limbic brain, which is a separate entity from the rational and decision-making neo-cortex. Adults generally have a developed pre-frontal cortex that allows them to bring strong emotions such as frustration and anger under control.
IBM describes this as the “inner voice of reason” that enables adults to keep strong emotional impulses in check.
However, the neo-cortex in children is still developing, which results in harnessing frustration and emotional outbursts. Generally being able to “do the right thing” is made more difficult for a child.
It says the AI toys may help, as usually an adult parent-figure who is monitoring the child’s behaviour is able to provide repeated cues and admonitions for behavioural change in these situations.
However, there are a number of disadvantages – like parental figures may not always be present or even when parental figures are present they may not want to be perceived as incessantly “nagging” the child.
The toys will work, says IBM, as they will constantly be there to monitor the child. However, we’re not sure Channel 4 will be too happy as they could effectively do it out of a cheap re-run schedule.
Berkley boffins have created an artificial intelligence program which can defeat a human Starcraft champion.
Dubbed Overmind, the software knocked the socks off Oriol Vinyals, who is a one-time World Cyber Games competitor. He managed to get to number 1 in Spain and was ranked in the top 16 players in Europe.
Overmind is Berkley’s entry into the 2010 StarCraft AI Competition.
StarCraft is one of the most popular games ever and professional StarCraft players are celebrities with six-figure contracts and their games broadcast live on national TV.
It is also a challenge for artificial intelligence because a successful StarCraft AI agent must attempt to solve a number of problems.
Writing in Ars Technica one of the boffins Haomiao Huang said that the complexity makes StarCraft an ideal environment for conducting AI research.
Events unfold in real-time and players’ orders are carried out immediately. Resources have to be gathered so fighting units can be produced and commanded into battle. The map is shrouded in fog-of-war, so enemy units and buildings are only visible when they’re near friendly buildings or units. In the past AI writers have looked at chess as a good way to test their skills because you have to look far into the future, go is worse because there are lots of pieces. With poker there’s uncertainty. However in StarCraft, you have all of these things going on simultaneously, and you have very little time to compute a solution.”
Humans get around the problem by practice and training, building up an extensive store of skills and expert knowledge. Getting this into an AI agent is nearly impossible.
Early attempts to beat humans proved to be a disaster. Humans were just too sneaky. Tactics meant little when a human would use cunning untought of techniques.
However now it seems that the Berkley team have mastered. Today Starcraft, tomorrow Starnet will will be controlling the universe. Watch it Sarah Conner.