The following is a complete list:
1. 3D metal printing
3D printing has been around for decades, but so far, 3D printing is still primarily used to make one-off prototypes, and if you want to use materials other than plastic, such as metal, for printing, it is not only costly but also speed very slow.
But at the moment, the cost of 3D metal printing is getting lower and lower, and has gradually become a way to make real parts, which, if widely adopted, will change the way mass production is done.
In the short term, manufacturers will no longer need to maintain a large inventory of products such as auto parts, print directly when needed.
In the long run, large factories that produce small parts in small quantities may be replaced by smaller ones that are better adapted to the changing needs of their customers.
3D metal printing can produce lighter weight, higher strength parts, complex shapes that are difficult to achieve with traditional methods, and more precise control of the metal microstructure. In 2017, a study from Lawrence Lvmo National Laboratory Staff have announced that they have developed a 3D printing method that produces stainless steel up to twice the strength of traditional processes.
Also in 2017, Boston-based 3D printing company Markforged unveiled its first 3D metal printer for less than $ 100,000.
Another Boston startup, Desktop Metal, launched a prototype metal printing machine in 2017. The company plans to sell a large machine for actual product production 100 times faster than traditional metal-based printing.
Printing of metal parts is now even easier. Desktop Metal's current software is designed to create designs that are suitable for 3D printing. Users simply tell the software which objects they wish to print and the software generates a computer model suitable for printing.
General Electric, which has long been using 3D printing technology in its aerospace products, is testing new metal-based printers at a fraction of the cost of large-format parts, and the company plans to sell the printer by 2018.
Artificial embryo
The researchers carefully placed the cells on a three-dimensional scaffold and observed the cells that were in communication with each other and arranged into the shape of a few days old mouse embryo.
Team leader Magdelena Zernicka-Goetz said: 'We know that the great potential of stem cells is a lot like magic, but we do not realize that they are so beautiful and perfect Self-organization.
Zernica-Gates said the 'synthetic' embryo may not develop into a mouse, but nevertheless it still shows that we can grow mammals without the need for egg cells.
This is not the goal of Zernijkar-Gates, who wants to study how early embryonic cells play a special role, she said, and the next step is to use human stem cells to grow artificial embryos. The University of Michigan and Rockefeller University are also working on this Aspects of research.
Artificial intelligence synthetic human embryos will bring scientists the gospel to help them understand every part of the body's early development, and since these embryos come from easy-to-manipulate stem cells, there are a variety of tools available in the lab, such as gene editing tools, Research in their growth.
However, artificial embryos pose an ethical problem. What happens if such embryos are indistinguishable from true embryos? How long can they grow in the laboratory before they feel pain? According to bioethicists, in science Before the race starts, we need to address these issues first.
3 perception of the city
The project was announced in October 2017 and construction will begin in 2019. From New York City, Alphabet's Sidewalk Labs is working with the Canadian government on the project, which is scheduled to take place in the waterfront industrial area of Toronto.
One of the goals of the project is to use a vast network of sensors to gather data on air quality, noise levels, population activities and more, and then to guide design, policy and technology decisions.
Such a plan requires that all vehicles be autopilot and enter the shared travel platform. Robots engage in trivial tasks such as delivering mail. Sidewalk Labs said it will open up access to the company's software and systems to allow other companies to base on it On the development of services, just as third-party development and application of smart phones.
The company plans to closely monitor public infrastructure, raising concerns about data management and privacy, but Sidewalk Labs said it could be lifted by working with communities and local authorities.
Rit Aggarwala, head of urban systems planning at Sidewalk Labs, said: 'What we did for Quayside was that the project was not only ambitious but also has a human touch.' That may help Quayside Avoid repeating the mistakes of previous smart city plans.
Waterfront Toronto, the government agency responsible for the development of Quayside, said other North American cities are also in contact with Sidewalk Labs in hopes of becoming the next city to partner with, said Will Fleissig, CEO of San Francisco, Denver, Los Angeles And Boston have called to get the introduction.
4. Artificial Intelligence for Everyone
So what's the solution? Cloud-based machine learning tools are bringing artificial intelligence to a wider audience, and so far, Amazon AWS is a leader in cloud-based artificial intelligence. Google is relying on the open-source artificial intelligence library TensorFlow to Amazon Challenges. Recently, Google also announced Cloud AutoML, a set of pre-trained systems that make artificial intelligence easier to use.
Microsoft also has Azure, a cloud computing platform with integrated artificial intelligence, and Microsoft has partnered with Amazon to provide Gluon, an open-source, deep learning library designed primarily for developing neural networks that make neural networks as easy to develop as mobile applications.
It is not yet clear which company will become a leader in cloud-based AI, but for winners this means huge business opportunities.
If the AI revolution permeates all walks of life, then these products will be an essential element.
At present, artificial intelligence is mainly used in the technology industry, where artificial intelligence creates efficiency and brings new products and services, but many other enterprises and industries also try to utilize artificial intelligence.If the pharmaceutical, manufacturing and energy, etc. Industry can also fully deploy this technology, then productivity will be greatly improved, the whole industry will be revolution.
However, most businesses still lack enough talent to figure out how to take advantage of cloud-based artificial intelligence, so Amazon and Google are also providing consulting services, and once cloud computing popularizes technology to all, the true AI revolution will kick-start .
5. antagonistic neural network
The problem is that imagination is needed to create new things, and imagination is not good at artificial intelligence.
In 2014, Ian Goodfellow, a PhD student at the University of Montreal, first came up with this solution in a bar's academic debate called the GAN. GAN lets two nerves The web confronts each other in the digital version of Cats and Gobbles.
Both networks use the same dataset to train, one of which is called 'Builder' and the task is to use the image you see to create different versions of someone with 3 hands, for example, while another is called 'Discriminator' , The task is to identify whether the image you see is a fake image created by the generator.
Through this process, the generator will be very good at generating images, resulting in the discriminator can not determine what is the real image, which is false. In essence, the generator is trained to identify and make the seemingly real image.
For the past 10 years, GAN has become one of the most promising areas of artificial intelligence, helping machines generate eye-catching results.
For example, Nvidia researchers provided GAN with a large number of celebrity photos and subsequently created hundreds of non-existent avatars, while another team of researchers generated A pseudograph that resembles Van Gogh's work.Furthermore, GAN can reimagine the picture in different ways, such as turning the sunny road into a snowy road or turning the horse into a zebra.
The result is not always perfect: GAN might put two handlebars on the bike or put the eyebrows on the wrong face, but since the resulting images and sounds tend to be very real, In the sense, GAN has begun to understand the underlying structure of the world seen and heard, which means in addition to imagination, artificial intelligence can gain more independent ability to understand the world you see.
6. Pakistani fish earplugs
One wearing a headset and the other holding a cellphone. The wearer of the earplug speaks in his or her own language - the default language is English - the App app translates the sentence, spreads over the phone and plays it out loud. People respond; the answer is translated, spread to the headphones to play.
Google Translate already has a conversation feature, and its iOS and Android apps allow two users to talk and the conversation will be automatically recognized and translated, but the background noise will make it difficult for the app to understand people's conversations or know when people stop Talk, when to start translating
Pixel Buds circumvents these problems by allowing the user to hold the right earplug while speaking, and to differentiate the interaction between the mobile phone and the earphone, to control the microphone artificially and to keep the caller in touch without having to take the phone .
Pixel Buds are poorly designed and widely criticized, and they look foolish and may not fit your ears or hard-wired to your phone.
However, cumbersome hardware is not difficult to figure out. Pixel Buds shows great promise of real-time understanding of communication between languages, and there is no need for fish anymore.
7. Zero carbon emissions of natural gas
A pilot power plant is testing a technology that will make the dream of turning natural gas into clean energy a reality at the US natural gas and oil refining industrial hub outside Houston: The 50-megawatt project, called Net Power, , The organizers believe that they can generate electricity at a lower cost, at least comparable to that of a standard natural gas power plant, and that all the carbon dioxide emitted during the operation can be fully recovered.
If so, that means that the world has found a way to extract carbon-free energy from fossil fuels at reasonable prices. Natural gas power plants can increase or reduce production on demand, avoid the high capital cost of nuclear power, To avoid instability in the supply of renewable energy.
Partners in Net Power include 8 Rivers Capital and Exelon Generation, a technology developer, and CB & I, an energy construction company that is currently in the factory-building phase and has already begun preliminary testing and plans to announce early versions in the coming months evaluation result.
The plant places the carbon dioxide emitted by the burning natural gas under high pressure and high temperature and uses the supercritical carbon dioxide it produces as a 'working fluid' to drive a special turbine. Most of the carbon dioxide can be recycled continuously and the rest can be recovered cheaply.
A key part of reducing costs is selling carbon dioxide. Today, the main use of carbon dioxide is to help extract oil from wells, which is a limited market and not a particularly environmentally-friendly market. However, the 'net electricity' project ultimately hopes to see The demand for cement in the manufacture and manufacture of plastics and other carbon-based materials is growing.
'Net electricity' does not solve all the problems of natural gas, especially in mining, but as long as we are using natural gas, it should be used as cleanly as possible. Of all the clean energy technologies under development, 'net electricity' It is one of the most promising reductions in carbon emissions.
Perfect online privacy
The tool is an emerging cryptographic protocol called zero-knowledge proof.Although researchers have been studying for decades for this purpose, interest has only ballooned in the past year, thanks in part to The public obsession with cryptocurrency, most of which is privately owned.
Much of the technology for zero-knowledge proof has benefited from Zcash, a digital currency introduced in late 2016. Developers at Zcash used cutting-edge encryption technology, called zk-SNARK, that allows users to trade anonymously.
In Bitcoin and most other public blockchain systems, this is generally not possible, and in those systems, everyone can see the content of the transaction.Although the transactions are theoretically anonymous, they can be combined with other data Together, tracking and even identifying users, Vitalik Buterin, the creator of Ethereum, the second-largest blockchain network in the world, described zk-SNARKs as an 'absolute change' Game rules of technology '.
For banks, this could be a way to use blockchain in payment systems without sacrificing customer privacy. Last year, JPMorgan Chase joined in its blockchain-based payment system zk-SNARKs.
Despite its promising future, zk-SNARKs are computationally cumbersome and slow, and require a so-called 'trusted setup' to create encryption keys that could compromise the entire system once the key falls into the wrong hands, Researchers are looking for an alternative to deploying zero-knowledge proofs more efficiently and without the need for a key.
9. Genetic Prediction
The sudden appearance of these report cards is due to the dramatic advances in genetic research, some of which involve more than one million people.
It turns out that the most common diseases and many behaviors and characteristics, including intelligence, are not caused by one or a few genes, but are the result of the synergies of many genes. Using the data from the acquired genes, scientists are creating so-called 'Multiple gene risk score' mechanism.
Although new DNA tests provide only probabilities rather than diagnoses, they can greatly benefit medicine, for example, if a woman at high risk for breast cancer increases the number of mammograms and a low-risk woman reduces the number of mammograms, These tests may find more real cancer and reduce false alarms.
Pharmaceutical companies can also apply these ratings to clinical trials of prophylactic drugs for diseases such as Alzheimer's or heart disease, etc. By picking out more vulnerable volunteers, they can test the drug's effect more accurately.
The problem is that these predictions are far from perfect.Who would like to know if they may have Alzheimer's disease? What if people with low cancer risk score postpone screening and then develop cancer?
Multi-gene scores are also debated because they predict any trait, not just disease, for example, they now test IQ tests at about 10% accuracy. As the scores rise, DNA IQ predictions are likely to become However, how do parents and educators use this information?
For behavioral geneticist Eric Turk heimer, the genetic data is mixed, making the new technology 'both exciting and worrying.'
Quantum leap of material
The new Quantum Computers have a bright future, but they also present a dilemma: They are far more computationally powerful than today's machines, powerful enough to be unimaginable, but we have not figured out how to apply them.
One possible and tempting possibility is to design molecules precisely.
Chemists have dreamed of new proteins for more effective drugs, new electrolytes that produce better batteries, compounds that convert sunlight directly into liquid fuels, and more efficient solar cells.
We do not have these, because it's hard for molecules to model on traditional computers, and even in a relatively simple molecule trying to simulate electronic motion, the computational complexity far outweighs the power of today's computers.
But for the quantum computer this is a piece of cake because it does not use numbers that represent numbers 1 and 0 but the 'qubits' of the quantum system. Recently, IBM researchers used a quantum computer with 7 A qubit mimics a small molecule composed of three atoms.
When scientists create machines that have more qubits, they should be able to model larger and more interesting molecules precisely, and just as importantly, quantum algorithms will evolve better.