May 27, 2020
Enjoy these interesting stories from the scientific community.
*-- Artificial intelligence can predict a person's personality using only a selfie --*
A picture is worth a thousand words, but what about a selfie? According to a new study, clues to a person's personality are encoded in a selfie -- clues that can be gleaned by artificial intelligence.
Computer models, with only a selfie to go by, proved better in tests at predicting a person's personality than human raters.
Studies have shown links between facial features and traits such as aggressiveness and risk-taking. However, the latest neuroscience research suggests that the human brain doesn't process individual facial features. Instead, the brain processes faces holistically.
To study links between personality and holistic facial features, such as facial symmetry and height-to-length ratio, researchers in Russia turned to artificial intelligence.
A team of scientists trained a system of artificial neural networks to analyze photographs of human faces and make personality judgments. To do this, they used the results of previous studies that showed connections between certain facial characteristics and personality traits.
Scientists used their model to analyze the photos of 12,000 volunteers who filled out a personality survey and sent in a selfie. Participants rated themselves on a scale for openness, conscientiousness, extraversion, agreeableness and neuroticism, called the Big Five traits.
When analyzing the faces of two individuals, the computer model made a correct personality comparison -- predicting which persons reported themselves as more extroverted or dominant, for example -- 58 percent of the time, a value above that of chance.
The new research, published Friday in the journal Scientific Reports, showed conscientiousness was the easiest personality trait to recognize by looking at a person's face.
"We circumvented the reliability limitations of human raters by developing a neural network and training it on a large data set labelled with self-reported Big Five traits," researchers wrote in their paper.
Authors of the new paper suggest the artificial intelligence models developed for the study could be used to analyze online dating profiles.
"Given that partner personality and match between two personalities predict friendship formation, long-term relationship satisfaction, and the outcomes of dyadic interaction in unstructured settings, the aid of artificial intelligence in making partner choices could help individuals to achieve more satisfying interaction outcomes," researchers wrote.
*-- Australian scientists set Internet speed record using single optical chip --*
Scientists in Australia have developed a novel optical chip that promises to dramatically boost Internet speeds.
Using a single chip, researchers achieved the fastest Internet data speeds ever recorded, 44.2 terabits per second -- speeds capable of downloading a thousand HD movies in less than a second. Researchers detailed the record-setting performance in a paper published Friday in the journal Nature Communications.
The new chip is what is called soliton crystal micro-comb.
"It consists of a very low loss ring engineered to support pulses of light that stabilize into chains like atoms in a crystal," Arnan Mitchell, distinguished professor and director of the Micro Nano Research Facility at RMIT University, told UPI in an email. "This soliton crystal micro-comb is much easier to get into a stable state than other types of combs and is very stable once it is there."
The chip works by splitting the light running through already laid Internet fibers into 80 unique channels, a feat that previously required a cumbersome device featuring 80 lasers. Instead of 80 individual lasers, the micro-comb integrates the equivalent of hundreds of high-quality infrared lasers into a single chip.
For the test, scientists used nearly 50 miles of existing optical fibers in the Melbourne metropolitan area.
"We modulated information onto the individual comb lines using sophisticated phase and amplitude multiplexing, sent the information around the loop and measured it again on its return," Mitchell said.
Because the chip was made using fabrication techniques commonly used to make large quantities of commercial computer chips, researchers expect production of the new optical chip could be quickly scaled.
Mitchell and his colleagues envision the tiny chips being deployed at Internet data traffic points, intersections on the information highway where data for entire neighborhoods or small towns converge.
As more and more people get online -- and more and more people work from home in the age of COVID-19 -- the technology could help alleviate the slowdowns caused by increased Internet usage.
"We are interested to talk with companies who work on high-speed transmission systems to partner on integrating these photonic systems into commercial transceivers," Mitchell said.
In addition to boosting Internet speeds for residential users, the technology could be incorporated into a variety of devices, from medical instruments to self-driving cars.
"Our team uses similar micro-combs for many other applications where we can take advantage of the combs to perform hundreds of measurements simultaneously," Mitchell said. "Beyond high-speed communications, we are currently working on using combs for applications in sensing, including spectroscopy, inertial sensing with potential applications in satellite positioning, ultra-fast analysis of radar signals and even systems for reading point of care biosensors."