Yunqi Information: [ Click to view more industry information ]
Here you can find the first-hand Shangyun information in different industries, what are you waiting for, come soon!
Although artificial intelligence companies vigorously promote facial emotion recognition software, psychologists question that emotion recognition is only easier said than done.
Hundreds of faces flashed one by one on the screen, some of them stared at the eyes, some were deflated, and some had their eyes closed, mouths raised, and their mouths wide open. Seeing these faces, you must answer a simple question: Is this person experiencing orgasm or pain?
来源：Adapted from GL Archive/Alamy
In 2018, psychologist Rachael Jack and her colleagues recruited 80 people to do this test. The team from Glasgow University in the UK recruited these participants from the West and East Asia in order to study a long-standing hot question: Can facial expressions really convey emotions?
Researchers have asked subjects to read emotions from their faces for decades, including adults and children in different countries, and even indigenous people in remote areas. In the 1960s and 1970s, a well-known observational study conducted by American psychologist Paul Ekman found that people all over the world can accurately deduce the emotions behind facial expressions, which shows that the expressions of emotions are interlinked.
This view is basically unchallenged for a generation. However, the new generation of psychologists and cognitive scientists questioned after reviewing these data. Many researchers now believe that the actual situation is much more complicated, and facial expressions have very different meanings in different situations and cultures. For example, Jack ’s research found that although Westerners and East Asians have similar understandings of facial expressions expressing pain, they do not agree on which expression expresses pleasure.
As for Ekman's conclusion that the human face is a window of emotional expression, the differences between the researchers are growing. But this does not prevent commercial companies and governments from "paying for" his claims and applying them in ways that will change people's destiny. For example, in many judicial systems in the West, reading the defendant ’s emotions is part of a fair trial. Justice Anthony Kennedy of the Supreme Court of the United States once wrote in 1992 that this is necessary to "understand the minds and thoughts of criminals."
Ekman has designed a controversial training program for the United States Transportation Security Administration (TSA), which was launched in 2007, and its core is to interpret emotions. The project is called "Screening Passengers by Observation Techniques" (SPOT). The main purpose is to train TSA personnel to monitor dozens of suspicious signs that appear on passengers. These signs may reflect their anxiety, deception and fear. mood. The project has been widely questioned by scientists, members of Congress, and civil society organizations such as the American Civil Liberties Union. They accused this approach of being inaccurate and also brought racial prejudice.
The successive voices of doubt have not stopped the top technology companies. They believe that emotions are easy to detect, and some of them have developed emotion recognition software. Right now, these software are being tested or promoted, and the scope of application includes evaluating the matching of job applicants and positions, lie detection, making ads more attractive, and detecting a range of diseases from dementia to depression.
The valuation of this industry is as high as tens of billions of dollars. Technology giants such as Microsoft, IBM, and Amazon, as well as some more specialized companies (such as Affectiva in Boston and NeuroData Lab in Miami) have introduced algorithms to detect emotions through faces.
Researchers are still arguing about whether human faces can faithfully express and perceive emotions. Many experts also believe that it is too early to use computers to automate them, especially because this technology is potentially destructive. The AI Now Institute at New York University ’s research center called for a ban on the use of emotion recognition technology in sensitive settings, such as recruitment and law enforcement.
Aleix Martinez, a researcher at Ohio State University who is engaged in related research, said that human facial expressions are difficult to interpret, even for humans themselves. He said that considering this, combined with the current trend that everything can be automated, "we should be worried."
Superficial view
There are 43 muscles on the human face. They can stretch, lift, and twist, expressing dozens of different expressions. Although facial muscles can perform many actions, scientists have always believed that certain expressions correspond to certain emotions.
Those who hold this view include Darwin. In 1859 he published the field investigation masterpiece "The Origin of Species" can be called a textbook on observation. Another of his less influential works, "Emotional Expression of Humans and Animals" (1872), is quite dogmatic.
Darwin noticed that primate facial movements are somewhat similar to human expression of emotions (such as disgust and fear). He proposed that these expressions must have some adaptive function. For example, expressions related to disgust, such as pouting your mouth, wrinkling your nose, and squeezing your eyes, may have been originally intended to defend against harmful pathogens. It was only with the emergence of social behavior that these facial expressions began to play a role in communication.
Darwin's discourse on emotions included a large number of pendulum expressions, such as those who tried to imitate the painful subjects. Source: Alamy
The first batch of cross-cultural field studies conducted by Ekman in the 1960s supported Darwin's hypothesis. He has studied human expressions and perceptions of six key emotions (happy, sad, angry, scared, surprised, and disgusted) around the world, and even includes a remote tribe in New Guinea.
Ekman told Nature that he chose these six emotions out of practical considerations. He said that some emotions, such as shame and guilt, do not have explicit expressions. "The six emotions that I focus on are expressive, that is, they can be used as research objects."
Ekman believes that these early studies support the universal expression theory extended by Darwin's theory of evolution. Later research proved that some facial expressions have adaptive advantages.
Lisa Feldman Barrett, a psychologist at Northeastern University in Boston, said: "For a long time, people think that facial expressions are a mandatory action." In other words, our faces cannot hide our emotions. However, an obvious loophole in this assumption is that people can indeed forge emotions, or they can keep emotions out of their faces. Ekman scholars also admit that there is no so-called "gold standard" for the expression of each emotion.
More and more researchers have suggested that the range of expressions corresponding to emotions is so large that the concept of the gold standard is almost falling apart. They supported this view with a large roundup. A few years ago, the editor of the journal Psychological Science in the Public Interest invited some mutually exclusive authors to form a panel of experts to complete this review.
Barrett, who led the collaboration, said: "We do our best to abandon the foresight." They did not make assumptions in advance, but started directly from the data. She said: "When the views are not unified, we will look for new evidence." Finally, they read about 1,000 papers, and after two and a half years of research, they came to a very obvious conclusion: no evidence, or very There is little evidence to prove that people can speculate on someone's emotional state from various facial movements.
The emotions reflected by the face are very limited
These researchers even cite some studies that prove that facial movements are not related to internal emotions. Carlos Crivelli, a psychologist at the University of De Montfort in the United Kingdom, has studied the residents of the Trobriand Islands in Papua New Guinea, and he found no evidence to support Ekman's views. Crivelli's conclusion is that inferring the internal mental state from external performance is like weighing with a ruler.
Another reason for the lack of evidence to prove the universality of expressions is that the face only provides part of the information. Other information, such as body movements, personality, tone and changes in face, also play an important role in the process of recognizing and expressing emotions. Just like changes in mood affect blood flow, blood flow affects complexion. Martinez and colleagues found that people can discover the relationship between changes in face and emotions. And visual signals such as background can also provide clues for identifying emotional states.
From top left: Basketball player Zion Williamson celebrates a slam dunk; Mexican fans celebrate the World Cup group stage promotion; singer Adele won the Grammy Award in 2012; fans of Justin Bieber cried at a concert in Mexico City.
Complex emotions
Other researchers pointed out that the counterattack on Ekman's conclusion was a bit overdone. Ekman himself took it for granted. In 2014, in response to Barrett's criticism, he pointed out that there were a large number of studies supporting his previous conclusions, including studies proving that the face will spontaneously express expressions. Other studies have discovered the connection between facial expressions and the state of the brain and body. In his response, he said that these studies show that facial expressions reflect not only human emotions, but also patterns of neurophysiological activity (see go.nature.com/2pmrjkh). He said his opinion has not changed.
According to Jessica Tracy, a psychologist at the University of British Columbia in Canada, the evidence given by those who believe that Ekman's expression universalism is wrong is just a small set of counterexamples, and they are exaggerating.
She believes that even if different groups or cultures have slightly misunderstood anger expressions, they cannot overthrow the whole theory. Most people know that this is an angry face at first glance, and she cites an analysis of 100 studies. She said: "There is a lot of other evidence that most people in most cultures around the world think this expression is universal."
Tracy and three other psychologists believe that Barrett said in the literature review that they are rigidly corresponding to the six emotions and facial movements. This interpretation is a bit exaggerated. One of the authors, Disa Sauter of the University of Amsterdam, said: "I don't think there are other researchers in the field of emotional science who agree with her."
Sauter and Tracy believe that to interpret facial expressions requires a more complex classification of emotions. Researchers should not treat happiness as a single emotion, but continue to subdivide it; happiness also includes happiness, joy, pity, pride, etc. The expressions of these emotions may differ or overlap.
Some studies use computers to generate random expressions. In a study conducted by Rachael Jack in 2018, participants need to indicate how well each face fits their definition of pain or orgasm. Source: C. Chen et al./PNAS (CC by 4.0)
The core of this controversy is actually the definition of distinctiveness. In one study, participants needed to choose one of six emotional tags to describe the face they saw. Some researchers may think that if the probability of a certain expression being selected is greater than 20%, it means that this expression is more versatile.
Others feel that the 20% standard is too loose. Jack believes that Ekman ’s threshold is too low. She read Ekman ’s early papers during her PhD. She said, “I always go to my mentor and show him these 60s and 70s charts. There are huge differences. As of today, there is still no data to prove that the recognition of emotions is universal. "
Even if saliency is not considered, researchers have to face the problem of subjectivity: many studies need to label emotions in advance so that they can be compared after the experiment. Therefore, Barrett, Jack, and other scholars want to study emotions in a more objective way. Barrett is studying physiological indicators, and she hopes to use them to describe anger, fear, and pleasure.
Jack uses computer-generated expressions instead of facial photos to avoid being limited to the six most common emotions. Researchers also asked participants to classify faces themselves, or participants from different cultures to tag photos in their native language.
Silicon-based sentiment
Software companies avoid free association of algorithms. In general, artificial intelligence algorithms for emotion recognition need to learn millions of face images and hundreds of hours of video-each emotion is labeled, and then learn patterns from these materials. Affectiva said the company has trained the software with more than 7 million faces from 87 countries, and its emotion recognition accuracy rate has reached 90%.
The company declined to disclose the scientific basis behind the algorithm. Neurodata Lab realized the difference in facial expressions, but pointed out: "If someone is experiencing a certain emotion, the possibility of certain facial expressions will be higher than the random probability." And the company's algorithm uses exactly This law. Researchers who are not uniform in their opinions, no matter where they stand, are skeptical about this type of software. Whether they are worried about the data used by training algorithms, or believe that the field is still inconclusive.
Ekman said he had challenged these companies directly. He wrote to several companies, but refused to disclose the company's name, saying only that "they are the world's largest software companies," and asked them for evidence that their automation technology was effective, but no response was received. He said, "In my opinion, their theory is not supported by evidence."
Martinez said eclectically that automated emotion recognition may be able to represent the average emotional response of a group. Affectiva has sold software to marketing agencies and certain brands to help them predict the response of specific consumers to a product or marketing tool.
Even if this software is wrong, it will not have much impact, and the effect of at most ads is not as expected. However, the application of some algorithms may change people's fate, such as interviews and border inspections. Last year, Hungary, Latvia and Greece tried a passenger pre-screening system to analyze lies by analyzing facial micro-expressions.
To calm this emotional-expression debate, different research methods are needed. Barrett is often invited to show her research to technology companies, and just went to Microsoft not long ago. She believes that researchers need to practice Darwin's way of writing "Origin of Species": "observe, observe, re-observe." Observe how people in real life communicate information through the face and body, not just in the laboratory. Then use the machine to record and analyze images from real life.
Barrett believes that more data and analysis techniques, rather than reviewing old data and experiments, can help researchers gain new knowledge. For this science that she and other researchers seem to be untenable, many technology companies are eager to try. She has challenged these companies: "We have reached the cliff, and whether artificial intelligence companies will continue to use the research hypothesis that is full of loopholes. , Or do what should be done? "
[Yunqi Online Classroom] Every day, product technology experts share!
Course address: https://yqh.aliyun.com/liveJoin the community now, face-to-face with experts, and keep abreast of the latest developments in the course!
【Yunqi Online Classroom Community】https://c.tb.cn/F3.Z8gvnK
Original Published: 2020-04-09
author: Nature natural scientific
article from: " Academic headline " for information may concern " academic headline "
If you find that there is suspected plagiarism in this community, please send an email to: [email protected] to report and provide relevant evidence. Once verified, the community will immediately delete the suspected infringing content.
Original link
This article is the original content of the Yunqi community and may not be reproduced without permission.
