Showing posts with label AI. Show all posts
Showing posts with label AI. Show all posts

Thursday, July 12, 2012

Expressive face helps robot bridge 'uncanny valley'


A good-looking robot is seriously hard to find. Robots can be pretty, some even handsome, but as soon as they get too realistic they start to creep us out. A new system for helping robots to generate more realistic expressions might go some way to help.




Many years ago, robotcists realized that as you morph an abstract robot into a human you generate a peak of unease - the "uncanny valley principle" - that makes people feel uncomfortable when a robot looks realistic but not realistic enough. Some say it's because they remind us of a corpse. However, researchhas shown that if you manipulate the robotic images so that they are more attractive, you can bypass this feeling of unease.


To create a robot we are more likely to accept, life-like expressions are vital. That's why Nicole Lazzeri at the University of Pisa, Italy, and her colleagues have designed a "Hybrid Engine for Facial Expressions Synthesis" (HEFES) - a facial animation engine that gives realistic expressions to a humanoid robot called FACE.


FACE's appearance is modelled on one of the team's wives. "It's really realistic," says Lazzeri, who presented the work at BioRob in Rome last month. See for yourself in the video above.


To mimic the myriad expressions that facial muscles are capable of achieving, the team placed 32 motors around FACE's skull and upper torso that manipulate its polymer skin in the same way that real muscles do.


To create expressions they used a combination of motor movements based on the Facial Action Coding System(FACS) - a system created over 30 years ago which codes facial expressions in terms of anatomic muscle movements.


HEFES is used to control FACE's expressions. It is essentially a mathematical programme that creates an "emotional space" which a person can use to choose an expression for FACE that exists anywhere between one or more basic emotions, including anger, disgust, fear, happiness, sadness, and surprise. The algorithm then works out which motors need to be moved to create that expression or transition between two or more.


The team evaluated the accuracy of their expressions by asking five autistic and 15 non-autistic children to identify a set of expressions performed first by FACE and then by a psychologist. Both groups were able to identify happiness, anger and sadness but less able identify fear, disgust and surprise.


So is it more attractive? I'm not convinced. But FACE's ability to smoothly transition between one emotion and another is pretty remarkable. And not too creepy.





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , ,

Attempt To Convert Prof Hawking’s Brainwaves Into Speech


An American scientist is to unveil details of work on the brain patterns of Prof Stephen Hawking which he says could help safeguard the physicist’s ability to communicate.



Prof Philip Low said he eventually hoped to allow Prof Hawking to “write” words with his brain as an alternative to his current speech system which interprets cheek muscle movements. Prof Low said the innovation would avert the risk of locked-in syndrome. Intel is working on an alternative.


Prof Hawking was diagnosed with motor neurone disease in 1963. In the 1980s he was able to use slight thumb movements to move a computer cursor to write sentences. His condition later worsened and he had to switch to a system which detects movements in his right cheek through an infrared sensor attached to his glasses which measures changes in light.


Because the nerves in his face continue to deteriorate his rate of speech has slowed to about one word a minute prompting him to look for an alternative. The fear is that Prof Hawking could ultimately lose the ability to communicate by body movement, leaving his brain effectively “locked in” his body.


In 2011, he allowed Prof Low to scan his brain using the iBrain device developed by the Silicon Valley-based start-up Neurovigil.


Prof Hawking will not attend the consciousness conference in his home town of Cambridge where Prof Low intends to discuss his findings, but a spokesman told the BBC: “Professor Hawking is always interested in supporting research into new technologies to help him communicate.”


Decoding brainwaves


The iBrain is a headset that records brain waves through EEG (electroencephalograph) readings – electrical activity recorded from the user’s scalp. Prof Low said he had designed computer software which could analyze the data and detect high frequency signals that had previously been thought lost because of the skull.”


An analogy would be that as you walk away from a concert hall where there’s music from a range of instruments,” he told the BBC.”As you go further away you will stop hearing high frequency elements like the violin and viola, but still hear the trombone and the cello. Well, the further you are away from the brain the more you lose the high frequency patterns.





The iBrain device collects EEG data which it transfers to a computer



“What we have done is found them and teased them back using the algorithm so they can be used.” Prof Low said that when Prof Hawking had thought about moving his limbs this had produced a signal which could be detected once his algorithm had been applied to the EEG data.


He said this could act as an “on-off switch” and produce speech if a bridge was built to a similar system already used by the cheek detection system. Prof Low said further work needed to be done to see if his equipment could distinguish different types of thoughts – such as imagining moving a left hand and a right leg.


If it turns out that this is the case he said Prof Hawking could use different combinations to create different types of virtual gestures, speeding up the rate he could select words at. To establish whether this is the case, Prof Low plans trials with other patients in the US.


Intel’s effort


The US chipmaker Intel announced, in January, that it had also started work to create a new communication system for Prof Hawking after he had asked the firm’s co-founder, Gordon Moore, if it could help him.


It is attempting to develop new 3D facial gesture recognition software to speed up the rate at which Prof Hawking can write. “These gestures will control a new user interface that takes advantage of the multi-gesture vocabulary and advances in word prediction technologies,” a spokeswoman told the BBC.


“We are working closely with Professor Hawking to understand his needs and design the system accordingly.”





Intel began working with Prof Hawking after he wrote a letter to its co-founder Gordon Moore in 2011





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , , , , ,

Wednesday, July 4, 2012

AI systems could fight cyberbullying


Smart software could detect online bullying via a database system that can identify even the subtlest of abusive comments




"I have been bullied my entire life. About how I look like a whale and how im not pretty enough. I cant get boyfriends because i refuse to have sex until I am married. I just dont know what to do anymore...:\" - Samantha, 16



Pleas for help like this one appear on social media and internet forums every day, written by desperate teenagers who live their entire lives online. Knowing you're not alone can help. That's the idea behind new software that matches up such messages with similar posts from other worried teenagers, letting them know that what they're experiencing isn't unusual. It might also be possible to spot bullying behaviour as it happens online.


Recent high-profile cases have made cyberbullying front page news. In January, 15-year-old Amanda Diane Cummings died after jumping in front of a bus on Staten Island, New York. She'd been subjected to a campaign of bullying on Facebook by other pupils at her school. Last September, Jamey Rodemeyer, a 15-year-old boy from Buffalo, New York, killed himself after being teased online about his sexuality. The cases sparked lawmakers to push through legislation, passed by the state senate last week, that makes cyberbullying a crime.


To help tackle one part of the problem, Karthik Dinakar at the Massachusetts Institute of Technology and colleagues have been working on a project that analyses the posts written by teenagers on A Thin Line, a website run by MTV. The site encourages teenagers to post their problems anonymously and other teenagers leave comments giving advice. Many of the posts concern bullying and worries about sex.


Each of the website's 5500 posts were fed through an algorithm trained to recognize certain clusters of words and then categorize each post according to one or more of 30 themes, ranging from "duration of a relationship" to "using naked pictures of girlfriend". The words "boyf" "trust" "cheat" "break" "upset" in the same story might indicate the post was about a relationship ending, for example. Once a label was assigned, the algorithm picked another story on the site that covered the same themes.


"All these teenagers are still growing emotionally, and there's a tendency to think that their experience is singular to themselves," says Dinakar. "It can let them know that they are not alone in their plight."


The software was tested using a set of new stories written by volunteers, which it analyzed and matched with stories from the website. The volunteers rated the system very positively. They felt that the stories picked using the thematic algorithm were always a much closer match than those chosen using a basic algorithm that just matched keywords. The system was presented at a conference on social media in Dublin, Ireland, earlier this month. MTV now plans to start using it to match stories live on the site, so teenagers can read about those in a similar plight.


Can artificial intelligence also stop cyberbullying at its source? After Amanda Cummings died, her memorial Facebook page was filled with offensive comments, leaving her parents understandably distraught. So Dinakar is also developing software that will help spot online bullying as it happens.


Facebook has taken steps to stop cyberbullying, but it primarily relies on users flagging up comments as inappropriate.


To find less-obvious forms of abuse, Dinakar built software that compares online posts to an open-source database called ConceptNet. This is a network of phrases and words and the relationships between them that lets computers understand what humans are talking about. This way the system can work out what might be a bullying comment, even though it contains no abusive words. For example, it would know that: "Put on a wig and lipstick and be who you really are" aimed at a boy might be a negative comment on his sexuality, because ConceptNet knows that girls usually wear make-up, while boys do not.


The idea is that software like this could be integrated into a social network. If it spots patterns of bullying behavior, it may either flash up a box warning the bully, ban offending posts, or offer help and advice to the victim. Dinakar wants to combine his two projects to create a detector that can pick up even the subtlest of attacks, such as "liking" a negative Facebook status to make a nasty point, for example. The research is due to appear in the journal ACM Transactions on Interactive Intelligent Systemsin July.


Danah Boydof Microsoft Research in Cambridge, Massachusetts, says that although this kind of work won't solve the problem of online bullying, it will help to improve our understanding of what happens online.


"I'm glad that these researchers are working to identify different types of meanness and cruelty," she says. "I am very hopeful that these kinds of techniques will lead to a more holistic understanding of the problem."





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at 1 comment:
Labels: , , , , , , ,

Saturday, June 30, 2012

Artificial brain with billion connections

Google has done the unthinkable — the creation of an ‘artificial brain’ from 16,000 computer processors, with more than a billion connections.


The team led by Google’s Jeff Dean then fed it random images culled from 10 million YouTube videos — and let it ‘learn’ by itself.


Surprisingly, the machine focused in on cats. “We never told it during the training ‘this is a cat’,” said Dean. “It basically invented the concept of a cat.”


The revelation


“Contrary to what appears to be a widely-held intuition, our experimental results reveal that it is possible to train a face detector without having to label images as containing a face or not,” says the team.


“We also find that the same network is sensitive to other high-level concepts such as cat faces and human bodies.”


“Starting with these learned features, we trained our network to obtain 15.8 per cent accuracy in recognizing 20,000 object categories from Image Net, a leap of 70 per cent relative improvement over the previous state-of-the-art,” it said,Daily Mail reports.


The ‘brain’ was a creation of the company’s ‘blue sky ideas’ lab, Google X, reportedly located in Google’s Mountain View, California, headquarters — known as ‘the Googolplex.’


Engineers are free to work on projects such as connected fridges that order groceries when they run low — or even tableware that can connect to social networks.


Other Google engineers have reportedly researched ideas as far-out as elevators to space.





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , , , , ,

A wearable robot helps disabled people walk

“It’s a nice feeling to be upright, to walk and to have people at eye level,” said a beaming Peter Kossmehl at the Potsdam Rehabilitation Centre in Germany. The 40-year-old from the German state of Brandenburg had just tried out a bionic exoskeleton — a wearable, battery-powered robot — that enables paraplegics to take a few steps again.


The rehab centre is one of the first facilities in Europe to test the robot, called Ekso and made by Ekso Bionics. The California-based company introduced Ekso in the fall of 2011. Now it is to be tested worldwide on paraplegics, stroke and multiple sclerosis patients and other people with lower-extremity paralysis or weakness.


“In Germany, only patients in Aachen have tested the robot Ekso so far — that was a few weeks ago,” said company spokesman Bastian Schink. After them, eight people in Potsdam strapped on and tried out the approximately 23-kilogram exoskeleton. With the help of sensors in its foot units, weight shifts are converted into steps.


“I’d like to give my patients the opportunity to stand erect again as soon as possible,” said Bettina Quentin, director of physiotherapy at the rehab center south of Berlin. But Quentin, like many experts, warns against excessive expectations.


“People who function well with their wheelchair will always be faster in them than with the exoskeleton,” said Jan Schwab, head of spinal cord injury research at the Charite University Hospital’s Department of Experimental Neurology in Berlin. “The psychological benefits of a patient standing upright shouldn’t be underestimated, though.” The reactions of the Potsdam patients appeared to confirm this. “I’m not walking by myself,” he remarked. “It’s only passive walking.” Kossmehl, too, thinks more development work is needed. “But it’s just the right aid for the rehab center,” he said.


In the view of Ruediger Rupp, director of the Department of Experimental Neurorehabilitation in the Paraplegiology Clinic at Heidelberg University Hospital, the robot is no substitute for a wheelchair, especially considering that it is not suitable for all patients.


There are about 60,000 paraplegics in Germany, “of whom fewer than 10 per cent are candidates (for the robot),” he said. “That’s a very select group.” Someone who can hardly move his or her torso, for example, would have great difficulties with the robot.





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , , , , ,

Google Builds Artificial Brain Which Can Recognize A Cat

The Google X laboratory has invented some pretty cool stuff: refrigerators that can order groceries when your food runs low, elevators that can perhaps reach outer space, self-driving cars. So it’s no surprise that their most recent design is the most advanced, highest functioning, most awesome invention ever… a computer that likes watching YouTube cats?


Okay, it’s a bit more advanced than that. Several years ago, Google scientists began creating a neural network for machine learning. The technique Google X employed for this project is called the “deep learning,” a method defined by its massive scale. In layman’s terms, they connected 16,000 computer processors and let the network they created roam free on the Internet so as to simulate a human brain learning.


Stanford University computer scientist Andrew Y. Ng, led the Google team in feeding the neural network 10 million random digital images from YouTube videos. The machine was not “supervised,” i.e. it was not told what a cat is or what features a cat has; it simply looked at the data randomly fed to it. Ng found that there was a small part of the computer’s “brain” that taught itself to recognize felines. “It basically invented the concept of a cat,” Google fellow Jeff Dean told the New York Times.


So Google may have created a machine that can teach itself. But what Ng and his team have done is not as new as you may think. Over the years, as the scale of software simulations has grown, machine learning systems have advanced; last year, Microsoft scientists suggested that the “deep learning” technique could be used to build computer systems to understand human speech. This Google X machine is the cream of the crop—twice as accurate as any other machine before it. However, “it is worth noting that our network is still tiny compared to the human visual cortex,” the researchers wrote, “which is a million times larger in terms of the number of neurons and synapses.”


After “viewing” random pictures from random YouTube videos, the neural network created a digital image of a cat based on its “memory” of the shapes it saw in the images. The cat the computer created is not any specific cat, but what the computer imagines to be a cat. Plato had his Forms, and now Google has its computer-generated cat image.





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , , , , ,

Friday, June 29, 2012

AI system helps spot signs of copper cable theft

copper_cable_theft


The industrial-scale theft of copper telecommunications cables is a massive problem: in the UK alone, cable worth £770 million was stolen from overhead and buried telephone lines and railway signalling systems last year.


Now an artificially-intelligent sensing system is helping to prevent such thefts in the UK, the Adaptive and Resilient Systems workshop in London heard last week.


"It's proven absolutely fantastic," says Luke Beeson, general manager for cable theft prevention at British Telecom (BT). "Police in Kent, London and South Yorkshire have made arrests thanks to the system – and although cables are still being cut, fewer are being stolen because we are getting to the scene quicker. And that means people's phones are back on line faster."


Cable theft doesn't only cause extended internet, phone, cable TV and cash machine outages – it also shuts down railway networks for hours and, sometimes, days.



Rich rewards

The ill-gotten rewards can be high. Metal thieves need only cut a cable in a street cabinet while an accomplice a kilometer away does the same. Then they winch the cable onto a truck. That kilometer of cable can fetch £20,000 on the scrap metal market. Even fiber optic cable is not immune from theft-related outages: thieves mistake it for shielded copper and take it anyway, or they damage nearby high-capacity fibers while stealing copper, Beeson says.


The technological countermeasure adopted by BT is an AI algorithm installed in its national network monitoring center – a secure facility in a secret UK location – which it has rolled out over the past three months.


Called the Rapid Assessment BT Incident Tracker – RABIT – the chunk of code monitors all 120 million kilometers of cable on BT's phone network. RABIT is a real-time system based on a neural network that has been trained to sense the difference between a telecoms cable being severed and a cable that has gradually failed – perhaps due to corrosion, falling trees, water seeping in, or perhaps incursion by farming machinery. It does this by undertaking line tests and bandwidth measurements to home in on a telltale signature of a cable cut.


This event data is then immediately plugged into a geospatial incident monitoring system called Saturn, which gives BT security staff a national view of trouble spots on its networks, including cyber attacks. Saturn also uses AI routines to spot patterns.



Catch the culprits

"This lets us plot patterns of activity as the cable thieves move across the country, showing the hot spots and letting you predict where they'll go next," says Robert Ghanea-Hercock, chief security researcher at BT's lab in Ipswich, UK.


"The target response time we're aiming for is 15 minutes, which is what the police say they need to catch them before they take the cable," he adds. However, so as not to aid the bad guys, BT is not saying what the response time is right now.


The encouraging news from BT's use of the system has inspired the UK's national railway provider, Network Rail, to explore technical measures to prevent theft of lineside signalling cable. BT's system cannot be used on the railways because the types of signals carried are too different, says Rachel Lowe at Network Rail in York.


"We are looking at a number of measures, from CCTV to marking cables – and various detection systems from third-party security companies are being assessed," she says.


One measure that hasn't worked so far, she says, is acoustic sensors. "They get false alarms from foxes and rabbits."





Blowfish12@2012 blowfish12.tk Author: Sudharsun. P. R.



Related articles

Posted by at No comments:
Labels: , , , , ,
Subscribe to: Posts (Atom)

Infolinks In Text Ads