What if AI could interpret your imagination, turning images in your mind into reality? While that sounds like a detail in a science fiction, researchers have now accomplished exactly this. Researchers found that they could reconstruct highly accurate images from brain activity by using the popular stable diffusion (扩散) image generation model. The authors wrote that unlike previous studies, they didn’t need to train the AI models to create these images.
The researchers said that they first predicted a potential representation, which is a model of the image’s data, from fMRI (功能性磁共振成像) signals. Then, the model was processed and noise was added to it through the diffusion process. Finally, the researchers decoded (解码) text representations from fMRI signals and used them as input to produce a final constructed image.
The researchers wrote that a few studies have produced good reconstructions of images but it was only after training. This resulted in limitations because training complex models are challenging and there are not many samples to work with. Prior to this new study, no other researchers had tried using diffusion models for visual reconstruction.
In the past, we’ve seen other examples of how brainwaves and brain functions can create images. In 2014, a Shanghai - based artist Jody Xiong used sensors to connect sixteen people with disabilities to balloons of paint. The people would then use their thoughts to burst specific balloons and create their own paintings.
With the advancement of AI, more and more researchers have been testing the ways AI models can work with the human brain. In a January 2022 study, researchers at Radboud University in the Netherlands trained an AI network on fMRI data from 1, 050 unique faces and converted the brain imaging results into actual images. The study found that the AI was able to perform reconstruction. In the latest study, the researchers found that current diffusion models can now achieve high-quality visual reconstruction.
1. How did researchers make images from brain activity in the previous research?A.By using generation model. | B.By training AI models. |
C.By reconstructing mind. | D.By predicting people’s imagination. |
A.The fMRI signals | B.The diffusion text. |
C.The processed model. | D.The text representation. |
A.Their training work is very challenging. |
B.They’ve made training models less complex. |
C.Nobody has reconstructed images by brainwaves. |
D.They can reconstruct accurate images from brain activity. |
A.Few researchers have used AI to make reconstruct images. |
B.Scientists should do more to make high - quality reconstruction. |
C.The AI models should be improved to work with fMRI data. |
D.The AI contributes a lot to good visual reconstruction. |
相似题推荐
【推荐1】Next time you go on holiday, it's going to feel pretty different. You may have to wear a face mask. And that means getting your point across in an unfamiliar language—all while keeping your mouth covered—could become a bit of a problem.
That's where the C-mask, by Japanese company Donut Robotics, comes in. It's a smart, Internet-connected mouth covering that can make calls, raise the wearer’s voice, and most helpfully for travelers, translate conversations into eight different languages.
If need be, you simply place it over a standard face mask and connect it via Bluetooth to a smartphone. A special app will then allow you to record live conversations, turn them into text and translate them into English, Chinese, French, Korean, Thai, Indonesian, Spanish or Vietnamese (currently only from Japanese)—making it easy to understand what people are saying. It won't actually speak the translated phrases for you, but it does have a built-in amplifier(扩音器)to help you make yourself heard.
“We worked hard for years to develop a robot and we have used what we've researched to create a product that responds to how COVID-19 has reshaped society,'' Donut's CEO recently told reporters. The firm has already raised more than 28 million yen via crowdfunding to develop the product, which will go on sale for 3 ,980 yen in Japan.
There are all manner of fashionable, eco-friendly masks you can now get your hands on— check out these in London, Singapore and Barcelona. But in creating an affordable, high-tech mask, Donut could well be on to something.
1. The C-mask is developed to _______.A.seek fashion. | B.reshape society. |
C.replace cellphones. | D.promote communication. |
A.How the C-mask speaks. |
B.How the C-mask works. |
C.What advantages the C-mask brings. |
D.What languages the C-mask translates. |
A.take a look. | B.earn a living. |
C.make a difference. | D.issue a warning. |
A.Health. | B.Tourism. | C.Fashion. | D.Technology. |
【推荐2】AR surgical guide, developed by South Korean medical-technology startup SKIA Company, is assisting breast-cancer surgeries at Ewha hospital in Seoul.
When a tablet PC is held above a patient’s chest, CT images of the area the tablet is pointed at are displayed on the screen, showing the precise location of tumors (肿瘤) and tissues. It’s like looking into the body before any cuts have been made. The tablet can be moved around above the chest to see the CT images from different angles.
Traditionally, a doctor would look through a patient’s CT results on a computer monitor, and then make cuts of what they believe is the right depth at the desired location to carve out tumors. The new AR technology works as an assistant helping surgeons operate with greater accuracy.
“The AR technology acts like a navigation (导航) app for surgeries. It’s much easier to get to your destination when you have a digital guide,” says Lee Jun-woo, a breast-endocrine surgeon and oncologist (肿瘤学家) at Ewha hospital who has been testing SKIA’s product on patients. The technology wouldn’t be possible without the ability of the hospital’s private 5G network to send data in time.
Earlier this year, Ewha hospital took its first steps toward cooperative remote surgeries by connecting an operating room through its private 5G network with doctors in other locations, enabling them to virtually join in a surgery and exchange information. “As remote-surgery technologies mature and expand, cooperative surgeries like the ones being tested at Ewha hospital could become more common, while hospitals could also use the technology to train doctors in techniques such as robotic surgeries,” adds Mr. Lee.
Beyond the medical space, South Korea is working to speed up the adoption of private 5G networks in different industries. The government set aside 60 billion won, roughly $46.2 million, to support new private 5G projects across South Korea in 2022 and 2023.
1. What does AR surgical guide assist breast-cancer surgeries with?A.Passing through the body. | B.Displaying CT images. |
C.Connecting a tablet PC. | D.Showing exactly where to cut. |
A.To give an example. | B.To make a comparison. |
C.To provide evidence. | D.To clarify a definition. |
A.A tablet PC. | B.A helpful assistant. |
C.A navigation app. | D.A private 5G network. |
A.Cooperative remote surgeries have been well recognized. |
B.Globally, more industries will adopt private 5G networks. |
C.Hopefully, the 5G technology will benefit more than surgeons. |
D.Many hospitals have used the technology to train doctors. |
【推荐3】Lighting has come a long way since Tom Edison lit his first incandescent bulb (白炽灯泡) in the 1880s. LED bulbs are popping up everywhere now, on planes, car headlights, in your phone. And engineers are exploring more ways to use LEDs—everything from wireless data streaming to secure communication systems and in-flight networking.
Engineer Harald Haas, co-founder of pureLiFi, explains how LEDs can be used to transmit information, “LEDs have the property that we can change the light that comes out of an LED very, very quickly. That change in the brightness is what we exploit in order to encode data extremely fast, so that a receiver will then see these changes in the light intensity in a way a human eye would not be able to detect. Then we have algorithms (算法) to recover these changes and get back the data stream.”
There are many advantages to using LEDs to transmit information. For one thing, LEDs can communicate much faster than WiFi. What’s more, the visible light spectrum (频谱) is about ten thousand times larger than the radio spectrum. This would allow communication systems to not only use a spectrum that’s already been set up, but vast amount of free spectrum, which is in cars, in our LED lights at home, in streetlights and so on. “It’s ubiquitous. It’s already there,” Haas says.
LiFi would also be more secure than WiFi. Because light can’t go through walls, people would not be able to log on to LiFi networks in the same way that they’re able to log on to and eavesdrop on (窃听) ongoing WiFi communications. Haas argues that LiFi would also be available in places where communication is typically difficult when we can’t use radio.
One interesting application could be to use car headlights to communicate with other drivers on the road. “We can use these LEDs to transmit data from car to car. Normally you see the car in front of you, but if you were able to relay high-definition video from, say, three cars in front of you, you could see earlier what’s happening. This is a way we can enhance safety on our roads,” Haas says.
Haas and his coworkers foresee the LED light industry changing rapidly in the near future to include additional features. “That is where LiFi plays a key role. From home sensing of interior, you’d find out if people have fallen down and the way you would navigate (导航) indoors. So many, many more applications would be possible with light,” Haas says.
1. What is essential to enable LEDs to send signals?A.The research work by pureLiFi. | B.Quickly changing light of LEDs. |
C.Common use of LEDs in daily life. | D.Accurate analysis of light intensity. |
A.LiFi’s speed is superfast. | B.LED lights are free to use. |
C.LEDs are stable in quality. | D.Available spectrum is everywhere. |
A.advantage in security | B.mature application |
C.promising prospect | D.rapid upgrades |
【推荐1】An inability to stand on one leg for 10 seconds in later life is linked to nearly double the risk of death from any cause within the next decade, according to a new study. The simple balance test may be useful to be included in routine physical exams for people in middle and old age, the research, which was published in British Journal of Sports Medicine, suggested.
While aging leads to a decline in physical fitness, muscle strength and flexibility, balance tends to be reasonably well-preserved until a person’s 50s, when it starts to decline relatively rapidly, the research noted. Previous research has linked the inability to stand on one leg to a greater risk of falls and to cognitive decline.
The study involved 1,702 people aged 51 to 75 living in Brazil, who were asked to balance unsupported on one leg during an initial check. Researchers told the participants to place the front of the free foot behind the standing leg, keep their arms by their sides and eyes. Fixed straight ahead. Up to three attempts on either foot were permitted.
The study participants had an average age of 61 and two-thirds of them were men. Around 1 in 5 failed to balance on one leg for 10 seconds at the initial checkup. Researchers monitored the participants after the initial checkup for a period of seven years, during which 7% of the people died. The proportion (比例)of deaths among those who failed the test (17.5%) was significantly higher than deaths among those who were able to balance for 10 seconds(4.5%).
The research was observational and didn’t reveal cause and effect. The study didn’t look at any possible biological mechanisms that might explain the link between poor balance and longevity.
The study noted that, overall, those who failed the test were in poorer health, with greater proportion of suffering from obesity, heart disease and high blood pressure. Diabetes was also more common among those who failed to complete the test.
1. What did the participants take during the research?A.Vision examinations. | B.Medical examinations. |
C.Flexibility tests. | D.Balance tests. |
A.By calculating the participant numbers. |
B.By comparing the participants’ lifetime. |
C.By identifying the participants’ diseases. |
D.By monitoring the participants’ pressure. |
A.It had few participants. |
B.It didn’t last very long. |
C.It lacked in-depth study. |
D.It was the first research on balance. |
A.Balance determines a person’s lifetime |
B.The balance test should be promoted to people |
C.The ability to stand on one leg can reflect health levels |
D.Being able to stand for 10 seconds means being healthy |
【推荐2】“The failure to play is now a serious issue and it calls for action for change,” says Sir Ken Robinson, a leading expert in education, creativity and human development. This is the driving force behind Outdoor Classroom Day - a global teacherled campaign, supported by Dirt is Good, a company producing daily chemical products.
Outdoor Classroom Day, taking place on 17th May and 1st November this year, will see schools around the world swap the inside for the outside and take learning into the playground and beyond to make playtime a key part of the school day. This might involve using natural objects like stones to do sums, or going on an insect hunt to encourage curiosity. By now, Outdoor Classroom Day has grown from a grassroots movement to a global campaign that is expected to benefit five million children and over 40,000 schools from all around the world in 2018.
This is helping to change the trend that sees many schools selling up or building on their playgrounds and cutting back on playtime to make more room for academic studies, while at home children’s lives are increasingly filled with organized activities intended to help them learn. Today globally 61% of parents surveyed in the Dirt is Good Qualitative Study said that children don’t know how to play without using technology.
Outdoor Classroom Day is making playing time happen, with 22% of participating schools having increased their playtime since joining the campaign. 93% of teachers surveyed saw improvements in children’s creativity after playing outside, and 97% believe that time outdoors is necessary for children to reach their full potential.
Scientific studies show that real play - the active, physical, selfdirected play - is essential for children to develop key life skills that are not taught elsewhere. Few would question the value of developing creativity, leadership, resourcefulness, and curiosity.
1. What do children do on Outdoor Classroom Day?A.Have a day off and go playing anywhere. |
B.Have P. E.classes on the playgrounds. |
C.Learn and play by using natural things outside. |
D.Play on their own without the teachers’ guidance. |
A.Technology helps children learn better at home. |
B.Parents prefer to home school their children. |
C.Schools stop providing playgrounds. |
D.Adults ignore the importance of playtime. |
A.To introduce. | B.To persuade. |
C.To entertain. | D.To advertise. |
【推荐3】Later this month, Tate Britain will host its first Van Gogh exhibition since 1947. Here we’ve collected 4 of his best paintings before the show.
Potato Eaters
Potato Eaters was Van Gogh’s first major work and, despite a distinct lack of colour, it was one that he was most proud of. His aim was to represent the difficulties of country life, so the farmers are painted in earthy tones with rough faces and bony hands. He was more concerned with conveying his message than achieving technical perfection in this painting.
Irises
Van Gogh finished almost 142 pieces during his stay in the Saint Remy de Provence asylum (精神病院), where he was a voluntary patient from 1889 to 1890. This piece was started shortly after he arrived there. The hospital garden provided much inspiration for the artist during his stay there. Originally, the flowers in this work were painted purple but the red pigment (色素) faded over time and has now turned blue.
Cafe Terrace at Night
Despite being painted more than 130 years ago, this cafe still exists in France and has since been renamed the Cafe Van Gogh. This artwork marks the first time Van Gogh’s famous post-impressionistic star-filled sky was seen in a piece, and it was painted on the ground, in person, rather than from memory. While the artwork doesn’t bear his signature (签名), it’s widely known to be Van Gogh’s, as he mentioned Cafe Terrace at Night in a number of his letters.
Wheat Field With Cypresses
This piece was inspired by the view from Van Gogh’s asylum window and was finished on the spot. The work was the title of three paintings that offer an insight into his stay at the hospital. Van Gogh was so pleased with the summer landscape that he made three copies there, one of the same size — which hangs in the National Gallery and two smaller copies which he planned to gift to his mother and sister.
1. Which of the mentioned paintings will Van Gogh probably love best?A.Potato Eaters. | B.Irises. |
C.Cafe Terrace at Night. | D.Wheat Field With Cypresses. |
A.Blue. | B.Red. | C.Purple. | D.Green. |
A.In England. | B.In France. |
C.In the asylum. | D.In the memorial museum. |
【推荐1】Recently, there have been huge advances in Artificial Intelligence (AI) programs that create art. The programs, which create artworks based on short descriptions, have amazed many people. They have also raised difficult questions.
In April, NewsForKids.net reported on Dall-E 2, a tool from OpenAI. Dall-E uses AI to create incredible paintings, photos, and other artworks in many different styles based on simple descriptions. Since then, OpenAI has changed its rules to allow anyone to sign up and use the tool. But it’s still putting strong control on the art that can be created.
But now there are several other AI tools that work in similar ways. These new tools are free or cheap, and don’t have some of Dall-E 2’s limits. MIdjourney is a website that allows users to create images from short descriptions. Stable Diffusion is similar, but users can also run the program on their own computers.
The new tools are entertaining and powerful. They allow people to quickly see almost any idea they can imagine, even if they aren’t artists. For some creative workers, like those at companies that create ads or images, the tools allow them to quickly explore new ideas.
Many artists are upset that they’re now competing against computers for attention, prizes and jobs. It’s hard for a single artist to compete against all the art and photography that have ever appeared on the Internet. In August,Jason Allen won first place at the Colotsfo State Fair with his picture Theatre D’opera Spatial, which was created with Midjourney. That upset many people.
The images and the descriptions used to train these AI programs were taken from the Internet. The training images were without permission, which has raised questions about whether what these tools are doing is fair or legal.
The creators of training pictures haven’t given permission for their images to be used, but their skills and talents are a huge part of what makes the new tools so successful. Some artists’ groups say artists should get to decide if their work can be used by these tools.
1. What should we do to create artworks with Dall-E 2 or Midjourney?A.Upload some photos. | B.Draw simple pictures |
C.Offer brief descriptions. | D.Write a short article. |
A.They are concerned about them. | B.They are discontented with them. |
C.They accept them unwillingly. | D.They welcome them warmly. |
A.Jason Allen’s attracting attention. | B.Jason Allen’s creating a picture. |
C.Jason Allen’s using Midjourney. | D.Jason Allen’s winning the prize. |
A.Lifestyle. | B.Business. | C.Education. | D.Science. |
【推荐2】Can artificial intelligence uncover a liar? It sounds like science fiction, but such an AI system is possible. Rada Mihalcea, a professor of computer science and engineering at the University of Michigan, has worked on deceit (欺骗) detection for about a decade.
The first thing that researchers working on artificial intelligence and machine learning need is data. In the case of the work that Mlhalcea did, they began with videos from actual court cases. For example, a defendant speaking in a trial in which he is found to have committed a crime could provide an example of deceit; they also employed words from witnesses as either example of truthful or deceitful statements. Altogether, they analyzed 121 videos — about half represented dishonest statements, and half truthful. It was this data that they used to build machine learning deceit detectors that ultimately had between a 60 to 75 percent accuracy rate.
One thing the system noticed is the use of pronouns—people who are lying would tend to less often use the word “I” or “we”, Mihalcea explains. “Instead, people who are lying would more often use ‘you,’ ‘yours’.” Besides, someone telling a lie would use “stronger words” such as “absolutely,” and “very,” that reflect certainty, while interestingly, people telling the truth were more likely to use words such as “maybe” or “probably.” As for gestures, she points out that someone being deceitful would be more likely to look directly into the eyes of the person questioning them. They also tended to use both hands when gesturing.
However, Mlhalcea’s work is “far from perfection,” she admits. “As a researcher, we are content that we were able to get to 75 percent accuracy. But looked at another way, that’s an error rate of one in four. ” Ultimately, she sees technology like this as being assistive for people — it could indicate that it noticed something “unusual” in a speaker’s statement, and then perhaps have a person “investigate more. ”
1. How did researchers create the deceit detection using artificial intelligence?A.They interviewed criminals in court cases. |
B.They involved AI system in a real-life trial. |
C.They employed witnesses to identify dishonest criminals. |
D.They fed the system with both truthful and deceitful statements. |
A.Gesture with one hand. | B.Question others directly. |
C.Use pronouns repeatedly. | D.Speak with strong certainty. |
A.It completely lives up to her expectation. |
B.It can be used as an aid for human beings. |
C.It can replace human beings to uncover a lie. |
D.It successfully lowers the error rate of detection. |
A.The use of pronouns and gestures in lie |
B.Rada Mihalcea’s research on lie detection |
C.The limitations of current lie detection technology |
D.Accuracy rate of machine learning deceit detectors |
【推荐3】A team of urban planners and information scientists at Tsinghua University in China has found that an AI-based urban planning system could outperform human experts in creating urban planning designs. In their study, the group introduced the factors for the ideal urban plan and how well their AI did when tested.
For much of history, cities have been left to grow organically — immediate needs were determined and urban planners attempted to fulfill those needs. Such a random approach to urban growth has often led to less-favored results. In more recent times, city officials and other planners have attempted to take a more logical approach to the problem by creating designs that account for livability and pollution controls.
This has resulted in the development of urban planning as a science and the emergence of urban development professionals. This science can be complicated due to the increasing number of variables that must be accounted for as a new development expands. In this new effort, the research team attempted to apply AI to ease the burden.
In hope that residents of the planned community could travel from their homes to essential services within 15 minutes, the researchers trained the system with prior plans crafted by humans along with factors that are believed to be good design features such as the inclusion of parks, general greenery, bike paths and places for entertainment, thus improving their quality of life.
Once the system was built, the research group used it to first design a small-size city block. They would adjust the system for further improvements and gradually increased the size and studied its results.
It turned out that the plans the system developed were as good or better than those created by humans yet a huge reduction in time (from hours to seconds). The team concludes that their system is not meant to replace human urban planners, but to free them from the mundane (单调的) parts of urban planning, allowing them to focus on overall concepts. Hopefully, it would be an efficiency booster, if done right.
1. Why was urban planning often unsatisfactory in history?A.Urban growth was not well-organized. |
B.Urban planning experts were in shortage. |
C.City officials failed to meet residents’ needs. |
D.City officials ignored the factors for ideal urban plan. |
A.Large community and eco parks. |
B.Economic development and job opportunities. |
C.Smooth traffic and advanced green belt. |
D.Convenient transport and public services. |
A.It reduces the burden on human urban planners. |
B.It guarantees optimal results for urban development. |
C.It ensures faster completion of urban planning projects. |
D.It eliminates the need for any human involvement in planning. |
A.Dismissive. | B.Approving. | C.Critical. | D.Ambiguous. |