Remember The Jetsons’ robot maid, Rosie? Massachusetts Institute of Technology (MIT) researchers think her future real-life incarnations can learn a thing or two from Steve Carell and other sitcom stars.
MIT says a computer that binge-watched YouTube videos and TV shows such as The Office, The Big Bang Theory and Desperate Housewives learned how to predict whether the actors were about to hug, kiss, shake hands or slap high fives — advances that eventually could help the next generation of artificial intelligence function less clumsily.
“It could help a robot move more fluidly through your living space,” lead researcher Carl Vondrick said in an interview. “The robot won’t want to start pouring milk if it thinks you’re about to pull the glass away.”
Photo: AFP
Vondrick also sees potential healthcare applications.
“If you can predict that someone’s about to fall down or start a fire or hurt themselves, it might give you a few seconds’ advance notice to intervene,” he said.
The findings — two years in the making at MIT’s Computer Science and Artificial Intelligence Laboratory — are to be presented at the International Conference on Computer Vision and Pattern Recognition in Las Vegas from today through Friday.
Vondrick, a doctoral candidate focusing on computer vision and machine learning with grants from Google and the National Science Foundation, worked with MIT professor Antonio Torralba and Hamed Pirsiavash, now at the University of Maryland. The trio wanted to see if they could create an algorithm that could mimic a human being’s intuition in anticipating what will happen next after two people meet.
To refine what is known in artificial intelligence studies as “predictive vision,” they needed to expose their machine-learning system to video showing humans greeting one another.
Cue what Vondrick acknowledges were “random videos off YouTube.” Six hundred hours of them, to be precise.
The researchers downloaded the videos and converted them into visual representations — a sort of numerical interpretation of pixels on a screen that the algorithm could read and search for complex patterns.
They then showed the computer clips from TV sitcoms it had never seen before — interactions between The Big Bang Theory stars Jim Parsons (Sheldon Cooper) and Kaley Cuoco (Penny), for example — and asked the algorithm to predict one second later whether the two would hug, kiss, shake hands or high-five.
The computer got it right more than 43 percent of the time. That may not sound like much, but it is better than existing algorithms with a 36 percent success rate. Humans make the right call 71 percent of the time.
In a video trailer of the study that showed the algorithm blowing it on a clip from The Office, the researchers quipped: “So it’s not perfect ... still a long way to go.”
That likely will involve even more binge-watching. Six hundred hours of video sounds like a lot, but it is not really that much. By the time we are 10 years old, we have logged nearly 60,000 hours of waking-hours experience.
“Humans are really good at predicting the immediate future,” Pirsiavash, the team member now based in Baltimore, said on Wednesday. “To have robots interact with humans seamlessly, the robot should be able to reason about the immediate future of our actions.”
Martial Hebert, director of the robotics institute at Carnegie Mellon University in Pittsburgh, who was not involved in the MIT study, called it “an important work.”
“Some argue that prediction is a central part of [artificial] intelligence,” Hebert said. “If you have a robot that can predict, you can map a deeper and more complicated understanding of the environment around it.”
The researchers’ biggest relief? The computer did all the binge-watching.
“We never had to watch the videos,” Vondrick said.
MAJOR BENEFICIARY: The company benefits from TSMC’s advanced packaging scarcity, given robust demand for Nvidia AI chips, analysts said ASE Technology Holding Co (ASE, 日月光投控), the world’s biggest chip packaging and testing service provider, yesterday said it is raising its equipment capital expenditure budget by 10 percent this year to expand leading-edge and advanced packing and testing capacity amid strong artificial intelligence (AI) and high-performance computing chip demand. This is on top of the 40 to 50 percent annual increase in its capital spending budget to more than the US$1.7 billion to announced in February. About half of the equipment capital expenditure would be spent on leading-edge and advanced packaging and testing technology, the company said. ASE is considered by analysts
TRANSFORMATION: Taiwan is now home to the largest Google hardware research and development center outside of the US, thanks to the nation’s economic policies President Tsai Ing-wen (蔡英文) yesterday attended an event marking the opening of Google’s second hardware research and development (R&D) office in Taiwan, which was held at New Taipei City’s Banciao District (板橋). This signals Taiwan’s transformation into the world’s largest Google hardware research and development center outside of the US, validating the nation’s economic policy in the past eight years, she said. The “five plus two” innovative industries policy, “six core strategic industries” initiative and infrastructure projects have grown the national industry and established resilient supply chains that withstood the COVID-19 pandemic, Tsai said. Taiwan has improved investment conditions of the domestic economy
Huawei Technologies Co’s (華為) latest smartphones carry a version of the advanced made-in-China processor it revealed last year, results from an independent analysis showed. This underscored the Chinese company’s ability to sustain production of the controversial chip. The Pura 70 series unveiled last week sports the Kirin 9010 processor, research firm TechInsights found during a teardown of the device. This is a newer version of the Kirin 9000s, made by Semiconductor Manufacturing International Corp (SMIC, 中芯) for the Mate 60 Pro, which had alarmed officials in Washington who thought a 7-nanometer chip was beyond China’s capabilities. Huawei has enjoyed a resurgence since
purpose: Tesla’s CEO sought to meet senior Chinese officials to discuss the rollout of its ‘full self-driving’ software in China and approval to transfer data they had collected Tesla Inc CEO Elon Musk arrived in Beijing yesterday on an unannounced visit, where he is expected to meet senior officials to discuss the rollout of "full self-driving" (FSD) software and permission to transfer data overseas, according to a person with knowledge of the matter. Chinese state media reported that he met Premier Li Qiang (李強) in Beijing, during which Li told Musk that Tesla's development in China could be regarded as a successful example of US-China economic and trade cooperation. Musk confirmed his meeting with the premier yesterday with a post on social media platform X. "Honored to meet with Premier Li