MIT announces top 10 scientific achievements of 2017: facial recognition payment, thermal energy batteries, etc. are on the list

According to foreign media reports, MIT announced the top ten scientific and technological achievements of the year, including helping paralyzed patients walk again, "face recognition" as a means of payment, and the gradual rise of artificial intelligence... These scientific and technological achievements are closely related to the fields of economy, politics, and medicine. Some technologies are becoming more mature, while others will take some time to be realized. But in any case, you should always know about them.

　　1. Good news for paralyzed people

Estimated time to achieve: 10 to 15 years

　　Pictured is a brain-reading chip with electrodes attached.

　　The electrodes stimulate the spinal cord.

In recent years, experiments on controlling machines with the human brain have become common. Now scientists hope to go a step further and bring good news to paralyzed patients. Researchers plan to directly connect brain reading devices to electrical stimulation devices installed on the body to create a "neural bypass" that allows the brain to control the body's actions and give paralyzed patients the right to walk again.

A research team from Case Western Reserve University in the United States implanted two information recording devices in the brain of a high-level paraplegic patient. The device is made of silicon and is smaller than a postage stamp. The surface is covered with hundreds of metal detection wires as thin as a hair, which can "hear" the instructions issued by neurons. In addition, the research team also implanted more than 16 thin electrodes into the patient's hand and arm muscles to form a complete "neural bypass".

With the help of this system, the patient can slowly raise his arms, control the opening and closing of his hands, and even lift a cup of water. Scientists hope to apply this technology to multiple senses. In addition to treating paralysis, blindness, amnesia and other diseases may also benefit from it.

2. Self-driving trucks

Estimated time to achieve: 5 to 10 years

　　The driver can interrupt the autonomous driving mode by pressing the red button on the right side of the steering wheel.

　　Otto said it has no intention of having self-driving trucks completely replace human drivers for at least a decade.

Several companies have already started testing self-driving trucks. Although there are still many technical issues to be resolved, supporters say that self-driving is safer and cheaper. Compared with self-driving cars, self-driving trucks can generate greater economic benefits, but the technical challenges are also more difficult. In addition, such trucks have also caused a lot of controversy because they will affect people's jobs.

Otto, a US company, is working on developing self-driving trucks. The core of its technology lies in the lidar system installed on the truck, which uses lasers to obtain detailed information about the surrounding environment. The truck is equipped with a customized micro-supercomputer the size of a toast box, which is cooled by liquid and can quickly process a large amount of information and use it to control the actions of various parts of the truck. In addition, the car is also equipped with a button to interrupt the autonomous driving mode.

3. Facial payment technology

Estimated time to achieve: Already achieved

　　Company employees can enter corporate buildings by scanning their faces.

In recent years, computer facial recognition technology has made significant progress. A Chinese startup called Face++ is valued at billions of dollars. Several APP platforms, including Alipay and Didi, have adopted Face++'s facial recognition payment system.

Facial recognition technology has great potential in the security field. Public security agencies can use this technology to identify criminal suspects in videos recorded in public places.

Baidu is one of several Chinese companies working on facial recognition technology. Using a system developed by Baidu, passengers can pick up their train tickets by simply scanning their faces.

4. Practical quantum computers

Estimated time to achieve: 4 to 5 years

Delft University of Technology in the Netherlands is a world leader in quantum computer technology, and it may be possible to develop this technology to a practical stage within a few years, enabling breakthroughs in coding technology, materials science, medical research, and artificial intelligence.

Scientists have never stopped studying quantum computers, but for many years they have remained at the theoretical stage. However, in 2016, many theoretical ideas finally became reality. Companies such as Google, IBM, Intel, and Microsoft have also generously donated money to provide financial support for technology research and development.

The Delft University of Technology research team pointed out that the biggest challenge facing quantum computers is that quantum bits (the basic unit of quantum information) need to achieve quantum superposition and quantum entanglement at the same time, and they are easily affected by noise, vibration, electric fields and other factors, resulting in errors. The research team used the "quasiparticles" just discovered in 2012 to effectively improve the stability of quantum computers.

5.360-degree all-round camera technology

Estimated time to achieve: Already achieved

Most amateur panoramic cameras on the market today are just mediocre. But the most advanced 360-degree cameras can fully display the scene in front of the lens, making the viewer feel as if they are actually there. This technology has bright development prospects in fields such as medical education, sports, disaster relief, etc. It can even help the virtual reality industry.

360-degree cameras are inseparable from the rise of smartphones and other technological advances. For example, 360-degree cameras consume more energy and generate more heat, but the efficient chips used in smartphones can solve this problem well.

Most 360-degree cameras are not equipped with a display or viewfinder, but camera manufacturers have developed related apps to make it easier for users to select scenes and browse the photos they have taken.

6. Thermal solar cells

Estimated time to achieve: 10 to 15 years

Traditional solar panels are often bulky, expensive, and inefficient. But scientists at MIT have developed a new solar device that can greatly improve the efficiency of solar energy conversion. The device works by first converting sunlight into heat energy, and then converting the heat energy into light within the spectrum that can be used by solar cells. In this way, solar cells can absorb more energy and are much more efficient.

The method currently has some problems, such as it can only be used in a vacuum and is too expensive, but as its efficiency continues to improve, the device will become more and more worthwhile.

In addition, researchers are exploring another method of solar thermal photovoltaic conversion. Since thermal energy is easier to store than electrical energy, if too much thermal energy is generated, it can be stored and converted into electrical energy in rainy weather. If researchers can invent suitable thermal energy storage equipment and further improve the energy conversion efficiency, the system can provide us with a continuous supply of clean and cheap energy.

7. Second-generation gene therapy

Estimated time to achieve: Already achieved

For decades, researchers have never given up their pursuit of gene therapy. They plan to use modified viruses to deliver healthy genes to patients to replace genes that cause disease. But the research road has been full of setbacks. Now, scientists have finally solved the "obstacle" facing gene therapy, bringing hope of recovery to patients with severe combined immunodeficiency (SCID) and other genetic diseases.

Early gene therapies repeatedly failed because the gene delivery mechanism was not rational enough. In some experiments, the virus delivered the gene it carried to the wrong genome, causing the patient to develop cancer or organ failure, and ultimately die.

Today, researchers have found some viruses that are more efficient in delivering genes, effectively solving the above problems. However, the technology still faces several major challenges. When treating diseases such as SCID, fewer genes are involved, and scientists can use drugs accurately; but when dealing with diseases with unknown causes such as Alzheimer's disease, the technology is still a little lacking.

8. Human Cell Atlas

Estimated time to achieve: within 5 years

Biologists are preparing to launch a super project: using modern genetics and biological technology to accurately measure and analyze millions of cells in the human body in detail, and draw a "cell atlas" of the human body. This will provide biologists with more sophisticated and complex biological models and accelerate the development of new drugs.

There are as many as 37.2 trillion cells in the human body. To classify them, scientists from all over the world will give each cell a specific "molecular signature" and set a "zip code" for each cell.

Previous studies have shown that the human body contains about 300 different types of cells, but the real number is far more than that. To correctly classify them, scientists used three technologies. The first is molecular microfluidics technology; the second is the use of efficient sequencing instruments to identify active genes in single cells; the third is advanced labeling and staining technology, which can accurately locate which organ or tissue the cell is in based on its genetic activity.

9. IoT “Zombie Networks”

Estimated time to achieve: Already achieved

The picture shows the impact range of a large-scale "botnet" attack in the United States in 2016.

As early as 2000, hackers began to break into computers connected to the Internet and control them on a large scale from a central system, forming the so-called "botnet". Today, the increasing popularity of electronic devices that can connect to the Internet has provided hackers with opportunities to take advantage of.

The only defense is to only run secure software on connected devices, which would prevent botnets from ever appearing. But IoT devices were not designed with security in mind.

Using "botnets", hackers can interfere with the Internet advertising profit model and provide false information to advertisers; or steal the account information of IoT users and transfer it to criminal organizations. However, this technology is most often used to make certain websites unavailable, and political organizations often use this method to "ban" websites they don't like.

10. Reinforcement Learning

Estimated time to achieve: within one to two years

　　Reinforcement learning technology helped AlphaGo successfully defeat top Go player Lee Sedol.

AlphaGo, which became famous 2009 years ago, uses reinforcement learning technology. In addition to improving autonomous driving technology, reinforcement learning can also help robots quickly learn new knowledge, greatly promoting the development of artificial intelligence.

The concept of reinforcement learning originates from nature. More than a century ago, a psychologist conducted an experiment. He put several cats in a box and let them escape by pressing a lever. After accidentally pressing the lever, the cats gradually associated this behavior with "escaping from the cage" and escaped from the cage faster and faster.

The development of reinforcement learning technology was once fraught with obstacles and little success, but last year's victory of AlphaGo over Lee Sedol, a top Go player, marked a significant advance in the technology.

Reinforcement learning has been used in many fields, such as improving autonomous driving technology. However, this technology also faces many challenges, such as requiring a large amount of data to be effective and requiring multiple exercises to succeed.