Menu
Scientific Advances
Natural compound speeds bone growth
Astronomers spy massive stars in the making
Frog gene map a leap forward for humans
Sun-shy mums may raise MS risk in babies
Soft fossils provide new target for ET search
Mammoth blood brought back to life
Outdoor exercise can boost self esteem
Athletes on growth hormone 'sprint faster'
Countdown begins to 520 day 'Mars mission'
Warmer planet to stress humans: study
Plasma rocket to shorten space voyages
Vaccine may trigger infant epilepsy onset
'Fingerprinting' points to dusty Australia
The hole in the ozone layer: 25 years on
Humans interbred with Neanderthals: analysis
Herschel shows star formation is slowing
Washing hands makes tough choices easier
'Face-book' to measure pain in mice
Science gives clues to World Cup success
Human sigh acts as a reset button
Expert confirms Phar Lap arsenic theory
Dictionary blunder a matter of gravity
Warning on high-dose vitamin D
Calling mum makes you feel better
Computer program recognises online sarcasm
Researchers have developed a computer program capable of identifying online sarcasm, although it is a long way from becoming a prosthetic for people with poor social skills.

Oren Tsur, a computer scientist at The Hebrew University in Jerusalem is discussing his work at a meeting of the Association for the Advancement of Artificial Intelligence in Washington, D.C. this week.

Sarcasm is a useful way to blunt the impact of criticism but people can often miss it when it's delivered online, where there are no contextual hints or social cues.

To cut through the confusion, Tsur and colleagues developed a computer program that can identify sarcasm in online communities with an accuracy rate of about 80%.

While there is still a long way to go before computers will be able to understand all the subtleties of humour, the new work might, among other things, help companies sort through comments about their products to find out what customers really think.

Consider, for example, a website that allows users to post their opinions about the products it sells. One comment says, "The size of this camera is great. It fits right into my pocket." Another says, "The size of this camera is great. I need a porter to carry it."

"A typical summarisation system will conclude that people are very, very happy with the size and weight" of the camera, says Tsur. "Obviously, this is not the case."
Set off by a joke

Funny enough, Tsur's attempt to program a sarcasm detector started as a joke.

When he was a university freshman, he received an email that thanked him for his previous contributions to an annual conference and asked him to be that year's program chair.

The email was clearly intended for someone else, but Tsur thought it was funny, so he responded with what he thought was an obviously sarcastic tone. His reply was taken seriously.

"They allowed me to postpone the deadline for submission and asked me what I was working on," says Tsur.

"I wrote back that I was working on detecting irony in email. They didn't get that either."

Since then, Tsur's interest in sarcasm had turned serious.
Machine learning

Tsur has developed a computer program that uses a strategy called "machine learning."

To begin, he and colleagues fed the computer 80 sarcastic sentences and several hundred non-sarcastic sentences that they had plucked from Amazon user reviews.

Sarcastic comments included: "Trees died for this book?" and, for a smart phone: "All the features you want - too bad they don't work!"

The program analysed the sentences and created hundreds of patterns that it used to evaluate a total of 66,000 reviews for 120 products sold online.

Each review contained an average of 15 sentences.

One of the patterns it figured out, for example, was that sentences that start with "I guess" and end with an ellipsis are often, though not always, sarcastic.

To test how well the program was working, the researchers gave 200 of the same product reviews to three independent reviewers. Results showed about an 80% agreement between computer and humans.

Given a few million tweets, the program performed with a similar level of accuracy.
Far from perfect

The program's performance is still far from perfect, probably because sarcasm is such a complicated social construct, says Dr Katherine Rankin, a neuropsychologist at the University of California, San Francisco.

"That's about as good as a person with bad social skills would do," she says.

While such a program might be good enough to help rank reviews, she adds, "if the purpose of having a computer program recognise sarcasm is to be like prosthetic for people with poor social skills, I'm not entirely sure those people are really going to benefit .

To truly get interpret a comment like "Oh, I LOVE working on Saturdays," Rankin says, people usually need to know something about the context of a situation and the person who's talking.

Cues like eye rolling and a lilting tone of voice help. None of those are available in online communities.

"Our brains pick up complex social cues and process many subtle things," she says. "Computers are nowhere near getting there."

Print
Climate change impact on malaria questioned
Single lens glasses can help prevent falls
Movies manipulate our primal response
Luminescent sharks become invisible
Synthetic biology research gets a hearing
Source of ancient carbon 'burp' detected
Why the goddess of love is in a spin
Computer program recognises online sarcasm
New dinosaur had record-sized horns
Physicists solve missing neutrino mystery
Milk from grass-fed cows may be better
Crabs caught spying on rivals' love claws
Lifestyle may not boost breast cancer gene risk
'Trade-off' gene for plants discovered
Pacific islands growing, not sinking
Caffeine addicts get no real perk
Velvet worm's deadly slime revealed
SpaceX cleared for Florida lift-off
Cyborg rights 'need debating now'
Sunlight shines on silver technology
Mountain biking as risky as football, diving
Dusty simulations may reveal planets
Legal fight over breast cancer gene
Unions call for urgent nano information
Solar panel attraction deadly for insects
Meat eaters munched many ways: study
Snakes may be in decline worldwide
Dogs dumbed down by domestication
Menu
Fossil sheds new light on 'dino-bird'
DNA 'spiderbot' is on the prowl
GM cotton use increases fruit pest problem
Warming to kill off a fifth of all lizards
Super massive black hole given the boot
Ball lightning could be 'all in the mind'
Immune system could be used to test for TB
Mobile phone cancer link unclear, study
Teen brain wired to take risks
Synchrotron probes Egyptian beads
Argonauts 'gulp' air to swim freely
Space station gets a new room
'Digital genome' to protect dying data formats
Sweep yields leads for new malaria drugs
Researchers snap signs of illegal fishing
Spectrum reveals supernova surprise
Scientists create synthetic life
Eavesdropping a waste of energy
Star caught eating its offspring
Megafauna die-off may have cooled planet
Hepatitis C no longer 'death sentence'
Atoms bring quantum computing closer
Visualisation staves off constant craving
Experts debate homeopathy funding