New Stanford research finds computers are better judges of personality than friends and family

New research shows that a computer's analysis of data can better judge a person's psychological traits than family and friends.

Computers can judge personality traits far more precisely than ever believed, according to newly published research.

In fact, they might do so better than one's friends and colleagues. The study, published Jan. 12 and conducted jointly by researchers at Stanford University and the University of Cambridge, compares the ability of computers and people to make accurate judgments about our personalities. People's judgments were based on their familiarity with the judged individual, while the computer used digital signals – Facebook "likes."

The researchers were Michal Kosinski, co-lead author and a postdoctoral fellow at Stanford's Department of Computer Science; Wu Youyou, co-lead author and a doctoral student at the University of Cambridge; and David Stillwell, a researcher at the University of Cambridge.

According to Kosinski, the findings reveal that by mining a person's Facebook "likes," a computer was able to predict a person's personality more accurately than most of their friends and family. Only a person's spouse came close to matching the computer's results.

The computer predictions were based on which articles, videos, artists and other items the person had liked on Facebook. The idea was to see how closely a computer prediction could match the subject's own scores on the five most basic personality dimensions: openness, conscientiousness, extraversion, agreeableness and neuroticism.

The researchers noted, "This is an emphatic demonstration of the ability of a person's psychological traits to be discovered by an analysis of data, not requiring any person-to-person interaction. It shows that machines can get to know us better than we'd previously thought, a crucial step in interactions between people and computers."

Kosinski, a computational social scientist, pointed out that "the findings also suggest that in the future, computers could be able to infer our psychological traits and react accordingly, leading to the emergence of emotionally intelligent and socially skilled machines."

"In this context," he added, "the human-computer interactions depicted in science fiction films such as Her seem not to be beyond our reach."

He said the research advances previous work from the University of Cambridge in 2013 that showed that a variety of psychological and demographic characteristics could be "predicted with startling accuracy" through Facebook likes.

The study's methodology

In the new study, researchers collected personality self-ratings of 86,220 volunteers using a standard, 100-item long personality questionnaire. Human judges, including Facebook friends and family members, expressed their judgment of a subject's personality using a 10-item questionnaire. Computer-based personality judgments, based on their Facebook likes, were obtained for the participants.

The results showed that a computer could more accurately predict the subject's personality than a work colleague by analyzing just 10 likes; more than a friend or a roommate with 70; a family member with 150; and a spouse with 300 likes.

"Given that an average Facebook user has about 227 likes (and this number is growing steadily), artificial intelligence has a potential to know us better than our closest companions do," wrote Kosinski and his colleagues.

Why are machines better in judging personality than human beings?

Kosinski said that computers have a couple of key advantages over human beings in the area of personality analysis. Above all, they can retain and access large quantities of information, and analyze all this data through algorithms.

This provides the accuracy that the human mind has a hard time achieving due to a human tendency to give too much weight to one or two examples or to lapse into non-rational ways of thinking, the researchers wrote.

Nevertheless, the authors concede that the detection of some personality traits might be best left to human beings, such as "those (traits) without digital footprints and those depending on subtle cognition."

'Digital footprints'

Wu, co-lead author of the study, explains that the plot behind a movie like Her (released in 2013) becomes increasingly realistic. The film involves a man who strikes up a relationship with an advanced computer operating system that promises to be an intuitive entity in its own right.

"The ability to accurately assess psychological traits and states, using digital footprints of behavior, occupies an important milestone on the path toward more social human-computer interactions," said Wu.

Such data-driven decisions could improve people's lives, the researchers said. For example, recruiters could better match candidates with jobs based on their personality, and companies could better match products and services with consumers' personalities.

"The ability to judge personality is an essential component of social living – from day-to-day decisions to long-term plans such as whom to marry, trust, hire or elect as president," said Stillwell.

Dystopia concerns

The researchers acknowledge that this type of research may conjure up privacy concerns about online data mining and tracking the activities of users.

"A future with our habits being an open book may seem dystopian to those who worry about privacy," they wrote.

Kosinski said, "We hope that consumers, technology developers and policymakers will tackle those challenges by supporting privacy-protecting laws and technologies, and giving the users full control over their digital footprints."

In July, Kosinski will begin a new appointment as an assistant professor at Stanford Graduate School of Business.

Source: Stanford university

Tailored 'activity coaching' by smartphone

Tailored activity monitoring. Credit: Image courtesy of University of Twente

Today's smartphone user can obtain a lot of data about his or her health, thanks to built-in or separate sensors. Researcher Harm op den Akker of the University of Twente (CTIT Institute) now takes this health monitoring to a higher level. Using the system he developed, the smartphone also acts as an 'activity coach': it advices the user to walk or take a rest. In what way the user wants to be addressed, is typically something the system learns by itself. Op den Akker conducted his research at Roessingh Research and Development in Enschede. October 17, he defends his PhD-thesis.

The new telemedicine system was tested for three months, among a group of COPD patients -- a chronic lung disease. For these patients, physical activity is very important but it can also lead to an oppressed feeling and thus, to over-cautiousness. Using the coaching system of Van den Akker, the patients carry a small movement sensor and a smartphone. The system calculates if it is advisable to take a rest or, on the other hand, have a walk. The system is 'context aware': it looks at the time of day, the weather, the surroundings of the patient and determines if the time is right for taking some exercise.

Tone of voice

In addition, the system knows how the patient wants to be addressed. Some people don't mind an imperative tone of voice 'go for a 10 minutes' walk', others prefer a more friendly advice: 'what if you would take a walk in the park?' Op den Akker designed learning algorithms for this: the system learns the preferences of the user by itself. Future versions of the system may not use text messages anymore, but an 'avatar' on the screen, enabling interaction with the user as well. For this, Op den Akker has started starting cooperation with the Human Media Interaction group of the University of Twente.

Roessingh Research & Development (RRD) is the research department of Roessingh rehabilitation centre in Enschede, The Netherlands. RRD closely cooperates with the University of Twente in many projects. Op den Akker conducted his research at RRD and UT's CTIT Institute, under supervision of Hermie Hermens, Professor in Neuromuscular Control and Telemedicine. A spin-off company of the university, Inertia Technology, developed the movement sensor used in this project.

Op den Akker's PhD-thesis is titled 'Smart tailoring of real-time physical activity coaching systems'.

Source: University of Twente

Electric cars without drivers

Autonomous vehicle in the car-sharing operation: After the renter has called for the car, it navigates autonomously to the pick-up area. Credit: © Fraunhofer IPA

E-Mobile will park independently in the future and will also be able to find the next charging station without a driver. Researchers are working on electric cars that can travel short distances autonomously. On the basis of cost-effective sensors, they are developing a dynamic model that perceives the environmental situation.

Whoever got his driver's license twenty years ago and is back in a car for the first time is going to be rubbing his eyes in amazement. Electronic helpers warn of a possible collision when parking and keep the necessary distance to the car ahead during traffic. There are lane departure, crosswind, blind spot and high beam assistants, not to mention the anti-lock system. The car is taking over step by step in the cockpit. Researchers at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA are one step ahead: They are dedicated to automated driving and are working on the vehicles of tomorrow, which can drive through traffic without human assistance. In this process, the Stuttgart engineers are particularly keeping an eye on electric cars.

The specialty of the researchers at the IPA is the development of robots. In the institute building, there is a prototype that independently finds its way on its four wheels through unknown territory. The challenges that are to be mastered are similar to those for automated driving. Here, as well, sensors need to recognize the environment so that the vehicle can navigate around obstacles and find its goal. Why not take advantage of that experience and apply it to the car, say the engineers in Stuttgart. That is why, one and a half years ago, an interdisciplinary team of computer scientists, mathematicians, electrical engineers and mechatronics engineers launched the project Afkar (a German abbreviation for "autonomous driving and intelligent chassis concept for an all-electric vehicle").

In a first step, the electric car is intended to learn to find a parking space and to park without a scratch. The idea behind this is that the car should be able to recharge itself with electricity without human help. This would be particularly important for car-sharing. Imagine the following scenario: The driver easily parks the car in a properly equipped parking garage on any randomly available parking space. The car takes care of everything else itself. It communicates via a wireless interface with the charging station and the parking garage management. In this process, it provides information about its charge level and its location. If the battery is empty and a charging station is free, it maneuvers in the corresponding parking bay and is charged inductively, without a cable. Then it makes room for the next electric car and rolls to a free parking space. In this way, the few existing charging stations can be used effectively.

Necessary technology is available

"The technology needed for this scenario is already available," says Afkar project manager Benjamin Maidel. He is referring to the robots of the institute that find their way easily in a known environment, such as a factory floor. Rebuilding a similar car does not take a lot of effort. Many modern cars already have most of the sensors that are required for this. The data that these devices collect just have to be combined and interpreted accordingly so that they provide a picture of the environment. The Fraunhofer experts are currently developing the necessary technology with the help of complex simulation programs. Soon, they want to test the results in practice on a demonstration vehicle.

It becomes more difficult when a car is intended to move autonomously in traffic. This requires sensors that can look hundreds of meters ahead as well as software that can react to any unforeseen events, whether that's a building site, a thunderstorm or snow. Maidel and his team are focusing on cameras, ultrasound, radar and laser scanners that perceive the surrounding area up to a distance of 200 to 300 meters.

The Afkar group will first go with their test car to a cordoned-off test area. For public roads, a special permit is required. "Whether autonomous driving makes a breakthrough will be decided, along with the right price, by customer acceptance and the legal framework. For example, the liability for accidents has to be re-regulated. The technology will probably conquer the market step by step," says Maidel. The advantages are obvious -- particularly for car-sharing vehicles. Any customer could use his smart phone to call a car, which would then drive to the desired location. Car-sharing companies could utilize their fleets more fully than they do today.

Source: Fraunhofer-Gesellschaft

Robots take over inspection of ballast tanks on ships

Doctoral degree candidate Dian Borgerink of the Robotics and Mechatronics department at the University of Twente developed the build-on robotic arm for RoboShip, working together with the department's chairman, Stefano Stramigioli, and master's student Jop Huttenhuis.
Credit: Image courtesy of University of Twente

A new robot for inspecting ballast water tanks on board ships is being developed by a Dutch-German partnership including the University of Twente. The robot is able to move independently along rails built into the tanks. At the moment, people still carry out such inspections, with ships being brought into dry dock for the purpose. The costs can be as high as € 700,000 per inspection. The RoboShip project offers great advantages, not only in terms of cost but also in terms of safety.

RoboShip is an independent, intelligent robotic platform on rails for use within the shipping industry. The parties involved in the RoboShip project brought together a number of innovations in it. Imotect, for example, developed a smart, cost-efficient rail, while DFKI was responsible for an autonomous vehicle to run along the rails. Incas³ developed the sensors, Xsens developed the tank navigation system, and the University of Twente was responsible for ensuring that a thorough inspection of the ballast water tanks is carried out using the equipment. The Meyer Werft shipyard in Germany has now also joined the project and will be integrating RoboShip into the ships it builds.

Operators are able to determine the exact position of the robot within the ballast water tanks thanks to a magnetic field. Other developments include a simple way of building synthetic rails into ships, an improvement to the supply of energy through the rails and advanced communication and navigation systems.

Current inspections: dangerous and expensive

Ballast water tanks are either filled with seawater or are left empty. Seawater is extremely aggressive and attacks steel. This is why regular inspection from within is required. Current practice is for a group of six inspectors to carry out such inspections. In doing so, they run the risk of injury through falls or of breathing in noxious gases. The autonomous RoboShip inspection robot makes such risks a thing of the past. Using the robot also greatly improves the efficiency of inspections.

Inspectors evaluate the data transmitted on a screen outside of the ballast water tank, considerably reducing the overall time required for inspecting a ship. The time that liners need to spend in dock can also be considerably reduced as tank inspections can take place while the ship is still in operation. Any repairs required can then be scheduled in before the ship goes to dry dock. Having a ship in dock is, after all, an expensive business, with costs running into the hundreds of thousands of euros. In the future, when the energy supply through the rails has been further improved, the robot will be able to work on the ballast water tank's surfaces using a laser. It will then be able to remove paint residues, for example, and it will also become possible to clean and coat the tanks.

Robotic arm from Twente

Doctoral degree candidate Dian Borgerink of the Robotics and Mechatronics department at the University of Twente developed the build-on robotic arm for RoboShip, working together with the department's chairman, Stefano Stramigioli, and master's student Jop Huttenhuis. Borgerink also works for the INCAS³ research institute, which developed the sensor system on the arm.

"I have had the opportunity of seeing the inside of a freighter's ballast water tank," says Borgerink. "After a voyage, it is slippery with seaweed and is full of noxious gases. Tanks like these are almost inaccessible due to ribs, pipes and cables. Realising that people actually need to go into them to carry out inspection work was what motivated me to develop the robotic arm." While designing the robotic arm, Borgerink and his colleagues were presented with many challenges. The arm needed to be collapsible, lightweight and accurate. The research carried out has already been presented at IROS 2014 in Chicago, a leading conference for the robotics and automation sector.

The Roboship research is funded by the German-Dutch INTERREG IVA subsidy programme.

Source: University of Twente

Researchers develop a system to reconstruct grape clusters in 3D, assess quality

Antonio José Sánchez Salmerón, researcher at the Instituto ai2 of the UPV, explains that, today, grape classification is based on an inspection by a panel of experts, that award it score depending on a series of parameters that determine its quality. Credit: Image courtesy of Asociación RUVID

Researchers of the Universitat Politècnica de València (UPV) have developed software to help reconstruct grape clusters with three-dimensional computer vision techniques. The system helps to automatically assess different parameters that define the quality of the wine grape during harvest time.

During the work, the researchers of the UPV collaborated with the Research Centre of Vine and Wine related Sciences of the University of La Rioja, the Spanish National Research Council (CSIC, in Spanish) and the Government of La Rioja. The results of this work were released last September in the journal Food Control.

Antonio José Sánchez Salmerón, researcher at the Instituto ai2 of the UPV, explains that, today, grape classification is based on an inspection by a panel of experts, that award it score depending on a series of parameters that determine its quality. Moreover, different tests are performed in the laboratory in order to estimate the quantity of sugar, the pH, the total acidity and the phenolic quality.

"Among the factors that define the quality of a wine, one of the most important is the quality of the grape as the raw material, but this concept is difficult to assess, due to problems such as subjective parameters, the short period of time available in the field to do the analysis during harvest time, the lack of measuring instruments and their high price, as well as the mixing of good quality and bad quality grape in the trucks. The introduction of this 3D grape reconstruction system helps assess different quality parameters for a wine grape cluster avoiding these problems. One of these parameters is the average size of the grape, which is a very important factor as it establishes the ratio between the quantity of skin and pulp," explains the researcher.

"Increasing the objectivity and automating the grape quality monitoring tasks would be a technological breakthrough with regard to the traditional evaluation system of the grape, based on the knowledge of an expert, and it would have a great impact on the wine industry," adds Sánchez.

Source: Asociación RUVID

Ancient wisdom boosts sustainability of biotech cotton

The patchwork of 75 million acres of small farms in northern China includes insecticidal transgenic Bt cotton. Credit: Photo courtesy of Yidong Wu

Advocates of biotech crops and those who favor traditional farming practices such as crop diversity often seem worlds apart, but a new study shows that these two approaches can be compatible. An international team led by Chinese scientists and Bruce Tabashnik at the University of Arizona's College of Agriculture and Life Sciences discovered that the diverse patchwork of crops in northern China slowed adaptation to genetically engineered cotton by a wide-ranging insect pest. The results are published in the advance online edition of Nature Biotechnology.

Genetically engineered cotton, corn and soybean produce proteins from the widespread soil bacterium Bacillus thuringiensis, or Bt, that kill certain insect pests but are harmless to most other creatures including people. These environmentally friendly toxins have been used by organic growers in sprays for decades and by mainstream farmers in engineered Bt crops since 1996. Planted on a cumulative total of more than half a billion hectares worldwide during the past two decades, Bt crops can reduce use of broadly toxic insecticides and increase farmers' profits. However, rapid evolution of resistance to Bt toxins by some pests has reduced the benefits of this approach.

To delay resistance, farmers plant refuges of insect host plants that do not make Bt toxins, which allows survival of insects that are susceptible to the toxins. When refuges near Bt crops produce many susceptible insects, it reduces the chances that two resistant insects will mate and produce resistant offspring. In the United States, Australia and most other countries, farmers were required to plant refuges of non-Bt cotton near the first type of Bt cotton that was commercialized, which produces one Bt toxin named Cry1Ac. Planting such non-Bt cotton refuges is credited with preventing evolution of resistance to Bt cotton by pink bollworm (Pectinophora gossypiella) in Arizona for more than a decade.

Yet in China, the world's number one cotton producer, refuges of non-Bt cotton have not been required. The Chinese approach relies on the previously untested idea that refuges of non-Bt cotton are not needed there because the most damaging pest, the cotton bollworm (Helicoverpa armigera), feeds on many crops other than cotton that do not make Bt toxins, such as corn, soybean and peanuts. The results reported in the new study provide the first strong evidence that these "natural refuges" of non-Bt crops other than cotton delay evolution of pest resistance to Bt cotton.

Tabashnik used computer simulations to project the consequences of different assumptions about the effects of natural refuges in northern China. The simulations mimic the biology of the cotton bollworm and the planting patterns of the 10 million farmers in northern China from 2010 to 2013, where Bt cotton accounts for 98 percent of all cotton, but cotton represents only 10 percent of the area planted with crops eaten by the cotton bollworm.

"Because nearly all of the cotton is Bt cotton, the simulations without natural refuges predicted that resistant insects would increase from one percent of the population in 2010 to more than 98 percent by 2013," said Tabashnik, who heads the UA's Department of Entomology and also is a member of the UA's BIO5 Institute. "Conversely, resistance barely increased under the most optimistic scenario modeled, where each hectare of the 90 percent natural refuge was equivalent to a hectare of non-Bt cotton refuge."

In a third scenario, the researchers used field data on emerging cotton bollworms from different crops to adjust the contribution of each hectare of natural refuge relative to non-Bt cotton. These data were provided by co-author Kongming Wu of the Institute of Plant Protection in Beijing. By this method, the total natural refuge area was equivalent to a 56 percent non-Bt cotton refuge, and 4.9 percent of the insects were predicted to be resistant by 2013.

To distinguish between these possibilities, a team led by co-author Yidong Wu of China's Nanjing Agricultural University tracked resistance from 2010 to 2013 at 17 sites in six provinces of northern China. Insects were collected from the field and more than 70,000 larvae were tested in laboratory feeding experiments to determine if they were resistant. This extensive monitoring showed that the percentage of resistant insects increased from one percent of the population in 2010 to 5.5 percent in 2013.

The field data imply that the natural refuges of non-Bt crops other than cotton delayed resistance with an effect similar to that of a 56 percent non-Bt cotton refuge, just as the model predicted.

"Our results mean we are getting a better understanding of what is going on," Tabashnik said. "We'd like to encourage further documentation work to track these trends. The same kind of analysis could be applied in areas in the U.S. where the natural refuge strategy is used.

"Natural refuges help, but are not a permanent solution," he added. "The paper indicates that if the current trajectory continues, more than half of the cotton bollworm population in northern China will be resistant to Bt cotton in a few years."

To avoid this, the authors recommend switching to cotton that produces two or more Bt toxins and integrating Bt cotton with other control tactics, such as biological control by predators and parasites.

"The most important lesson is that we don't need to choose between biotechnology and traditional agriculture," Tabashnik said. "Instead, we can use the best practices from both approaches to maximize agricultural productivity and sustainability."

Source:University of Arizona

Getting Bot Besponders into Shape - Scientists are tackling one of the biggest barriers to the use of robots in emergency response: energy efficiency

Steve Buerger is leading a Sandia National Laboratories project to demonstrate how energy efficient biped walking robots could become. Increased efficiency could enable bots to operate for much longer periods of time without recharging batteries, an important factor in emergency situations. Credit: Photo by Randy Montoya

Sandia National Laboratories is tackling one of the biggest barriers to the use of robots in emergency response: energy efficiency.

Through a project supported by the Defense Advanced Research Projects Agency (DARPA), Sandia is developing technology that will dramatically improve the endurance of legged robots, helping them operate for long periods while performing the types of locomotion most relevant to disaster response scenarios.

One of Sandia's new robots that showcases this technology will be demonstrated at an exposition to be held in conjunction with the DARPA Robotics Challenge Finals next June.

As the finals draw closer, some of the most advanced robotics research and development organizations in the world are racing to develop emergency response robots that can complete a battery of tasks specified by DARPA. Competing robots will face degraded physical environments that simulate conditions likely to occur in a natural or human-made disaster. Many robots will walk on legs to allow them to negotiate challenging terrain.

Sandia's robots won't compete in the finals next June, but they could ultimately help the winning robots extend their battery life until their life-saving work is done.

"We'll demonstrate how energy efficient biped walking robots could become. Increased efficiency could allow robots similar to those used for the competition to operate for much longer periods of time without recharging batteries," said project lead Steve Buerger of Sandia's Intelligent Systems Control Dept.

Batteries need to last for emergency response robots

Battery life is an important concern in the usefulness of robots for emergency response.
"You can have the biggest, baddest, toughest robot on the planet, but if its battery life is 10 or 20 minutes, as many are right now, that robot cannot possibly function in an emergency situation, when lives are at stake," said Buerger.

The first robot Sandia is developing in support of the DARPA Challenge, is known as STEPPR for Sandia Transmission Efficient Prototype Promoting Research. It is a fully functional research platform that allows developers to try different joint-level mechanisms that function like elbows and knees to quantify how much energy is used.

Sandia's second robot, WANDERER for Walking Anthropomorphic Novelly Driven Efficient Robot for Emergency Response, will be a more optimized and better-packaged prototype.

Energy-efficient actuators key to testing

The key to the testing is Sandia's novel, energy-efficient actuators, which move the robots' joints. The actuation system uses efficient, brushless DC motors with very high torque-to-weight ratios, very efficient low-ratio transmissions and specially designed passive mechanisms customized for each joint to ensure energy efficiency.

"We take advantage of dynamic characteristics that are common to a wide variety of legged behaviors and add a set of 'support elements,' including springs and variable transmissions, that keep the motors operating at more efficient speed-torque conditions, reducing losses," Buerger said.

Electric motors are particularly inefficient when providing large torques at low speeds, for example, to a crouching robot, Buerger said. A simple support element, such as a spring, would provide torque, reducing the load on the motor.

"The support elements also allow robots to self-adjust when they change behaviors. When they change from level walking to uphill walking, for example, they can make subtle adjustments to their joint dynamics to optimize efficiency under the new condition," Buerger said.

Robots must adapt to the diverse kinds of conditions expected in emergency response scenarios.
"Certain legged robot designs are extremely efficient when walking on level ground, but function extremely inefficiently under other conditions or cannot walk over different types of terrains. Robots need an actuation system to enable efficient locomotion in many different conditions," Buerger said. "That is what the adjustable support elements can do."

Early testing has shown STEPPR to operate efficiently and quietly.

"Noise is lost energy, so being quiet goes hand-in-hand with being efficient. Most robots make a lot of noise, and that can be a major drawback for some applications," Buerger said.

Robots' electronics, certain software to be publicly released

STEPPR's and WANDERER's electronics and low-level software are being developed by the Open Source Robotics Foundation. The designs will be publicly released, allowing engineers and designers all over the world to take advantage of advances.

The Florida Institute for Human and Machine Cognition is developing energy-efficient walking control algorithms for both robots. The Massachusetts Institute of Technology and Globe Motors also are contributing to the project.

WATCH VIDEO



Source: Sandia National Laboratories