Author | Gong Xueyuan

The application of artificial intelligence technology may help scientists on the Blue Planet escape the pain of endless experiments, accelerate the discovery of major scientific theories, and elevate human civilization to a new level.

--Title

The potential of AI technology is obvious to all, but what AI can be used for in the future and how much change it will bring to human society is also testing our imagination.

Although AI technology is still in the early stages of development, its existing capabilities are sufficient to change many fields, especially those with a lot of data but can not be effectively utilized.

1 Artificial intelligence promotes theoretical breakthroughs in basic science

In fact, the research process in basic science fields such as materials, chemistry, and physics is full of "big data," and scientists can't get away from collecting, selecting, and analyzing data from design, experimentation, testing, and proof.

Data in these fields are often structured, high-quality, and labelable due to the laws of physics, chemistry, or mechanics.

Artificial Intelligence (AI) technology (machine learning algorithms) is good at finding the "hidden" cause and effect relationships in large amounts of data, and is able to quickly process structured data in scientific research, so it has gained a lot of attention from researchers.

Artificial intelligence is leading the "post-modernization" of basic scientific research by showing great advantages in research in fields such as materials, chemistry and physics.

Taking the field of physics as an example, the application of AI has brought unprecedented opportunities for research in particle physics and space physics. In order to search for the Higgs boson (God particle) and further understand the microscopic composition of matter, the European Center for Nuclear Research (CERN) led the development of the Large Hadron Collider (LHC).

The LHC is currently the world's largest particle gas pedal, and it can generate one million gigabytes (GB) of data per second, which is surprisingly comparable to the amount of data Facebook accumulates in an hour in a year.

Some researchers have come up with the idea of using specialized hardware and software, and machine-learning techniques to decide in real time what data needs to be saved and what can be discarded.

Machine-learning algorithms have been shown to make at least 70 percent of those decisions, and can greatly reduce the workload of human scientists.

While it's easier to focus on the commercialization of AI, the application of AI to basic scientific research is much more exciting.

Because the transformation of social productivity comes down to further breakthroughs in basic scientific research.

We may never return to the days of scientific "giants" like Newton, Maxwell and Einstein.

In those days, the "giants" could write simple, beautiful theorems on paper or design world-beating experiments with wisdom beyond their time.

There may not be many more opportunities for such great work, and in this era there is a greater need for work that uses large amounts of experimental data to obtain the truth.

The search for the origin of the universe, or the folding of protein molecules, has involved one group of scientists after another.

The application of AI technology may help scientists on the Blue Planet get rid of the pain of endless experiments, accelerate the discovery of major scientific theories, and elevate human civilization to a new level.

2 Artificial Intelligence to promote the rapid improvement of social productivity

Artificial Intelligence is undoubtedly the highest goal and ultimate vision of computer applications:

To completely liberate human beings from repetitive mechanical labor, and to allow people to engage in work that is truly in line with the level of human intelligence and full of creativity.

In the 60-year history of AI development, far-reaching applications such as machine translation, image recognition, voice assistants and personalized recommendations have been born, and people's lives have changed dramatically without their knowledge.

In the future, can the further extension of AI application scenarios bring about a great improvement in social productivity and lead humanity into a new era?

In order to explore this question, Wu Enda, who has served as an executive at Google and Baidu, set up a company based on solving the problem of AI transformation in 2017 Landing . ai.

Enda Wu announced the company's founding on his personal social media sites through a post and a video, saying he hoped AI could transform every aspect of human life, including food, clothing, housing and transportation, and free people from the mental drudgery of repetitive labor.

Landing, which means "landing" in Chinese, aims to help traditional companies use algorithms to reduce costs, improve quality management and eliminate supply chain bottlenecks.

So far, Landing . ai has chosen two landing areas, respectively, manufacturing and agriculture.

Landing . ai was the first to partner with manufacturing giant Foxconnect.

Landing . ai tries to help Foxconn move towards smart manufacturing, artificial intelligence and big data by utilizing technologies such as automated visual inspection, supervised learning and prediction to elevate the level of AI adoption in the manufacturing process.

Wu Enda believes that the impact of AI on manufacturing will be as powerful as the invention of electricity in the first place, and that AI technology is well suited to address some of the challenges currently facing the manufacturing industry, such as inconsistent quality and output, lack of flexibility in production line design, inability to keep up with capacity management, and rising production costs.

Concepts such as the industrial Internet, smart manufacturing and Industry 4.0 have taken hold, and traditional companies are transforming into intelligent, data-driven businesses, but how to apply the massive amounts of data acquired during the production process has become a new problem. landing . ai's cooperation with Foxconn may bring new insights to practitioners of traditional manufacturing.

Of course, the core competitiveness of the manufacturing industry also lies in the manufacturing itself, such as the precision of the lathe, the ability to control the temperature of the heat-treatment furnace, and so on, and the core competitiveness of the agriculture industry also lies in the agriculture itself, such as the breeding technology, genetically modified technology, and so on.

The main value of artificial intelligence technology is to improve decision-making ability, further improve production efficiency, and reduce the repetitive labor of human beings and other aspects, which is why artificial intelligence can be "empowered" by the reasons of various industries.

3 Artificial Intelligence will effectively improve the living space of human beings

Since the first industrial revolution, human activities have had an increasing impact on the natural world, and the growing demand for resources has led to huge changes in land use, increasingly serious pollution, and a sharp decline in biodiversity, making the living space of human beings increasingly harsh.

Entering the era of artificial intelligence, how to better utilize cutting-edge technologies such as big data and machine learning to empower environmental protection and green industries has become a focus of attention for governments, scientists, the public, as well as businesses.

In terms of energy utilization, Google's DeepMind is undoubtedly at the forefront.

Beginning in 2016, DeepMind introduced artificial intelligence tools to Google's data centers to help the tech giant save on energy bills.

DeepMind used a system of neural networks to recognize patterns to predict changes in power and used AI technology to manipulate computer servers and related cooling systems, successfully helping Google save 40 percent of its energy and boosting Google's overall energy efficiency by 15 percent.

After 2018, DeepMind has even extended its "tentacles" to the field of clean energy. As we all know, wind power is difficult to integrate into the grid and cannot be effectively utilized because of its volatility and unpredictability.

DeepMind uses data from weather forecasts, meteorological observations and other data to train neural network models that can provide wind forecasts 36 hours from now, thus making wind power on farms predictable.

Once wind power is predictable, power plants will have plenty of time to activate means of generating power that take longer to come online, complementing the wind. In this way, the problem of connecting wind power to the grid can be easily solved.

DeepMind's predicted wind power generation and actual power generation comparison

In terms of natural environmental protection, Microsoft's "Artificial Intelligence Planet Earth (AI of Earth)" has set an example for everyone.

Launched in July 2017, this program aims to protect and sustain the Earth and its natural resources with the help of cloud computing, IoT and AI technologies, providing support to individuals and organizations in the fields of water resources, agriculture, biodiversity, and climate change through grants, training, and in-depth cooperation.

For example, the "SilviaTerra" project enables forest monitoring by using Microsoft Azure, high-resolution satellite imagery, and U.S. Forest Service field data to train machine-learning models; the "WildMe The "WildMe" project uses computer vision and deep learning algorithms to identify endangered animals; and the "FarmBeats" project improves data collection in outdoor environments through sensors, drones, and other devices to improve the sustainability of agriculture.

In the first three industrial revolutions, scientific and technological advances brought great convenience to people's lives, but also brought difficult natural environmental problems such as climate change, biodiversity degradation, and atmospheric and oceanic pollution, which are gradually making the living environment worse for human beings.

On the surface, it seems that the natural environment of developed economies has improved, but this improvement is at the expense of transferring pollution and destroying the natural environment of developing countries, and the overall natural environment of the world is still not optimistic.

People have been pinning their hopes on future scientific and technological advances to solve the current natural environment problems, and the emergence of artificial intelligence technology has ignited this hope.

Once AI technology can accelerate the breakthrough of basic scientific theories, realize a significant increase in productivity, and effectively improve the living space of human beings, all the problems of development and the natural environment will be solved.

4 Conclusion

Standing in 2019 to look at artificial intelligence, can not help but feel a few chills. Artificial intelligence algorithms have no obvious breakthroughs, poor robustness, algorithmic black box and other problems are still prominent, part of the commercialization of landing is not as expected, some experts and scholars began to worry that artificial intelligence will usher in a new "winter".

But if we look back at AI in the future, all the current concerns will be just a hiccup.

Even now, the potential of AI technology is far from over.

The changes that AI is about to bring about will still be beyond most people's wildest dreams.

The disruption brought about by AI has been tangibly felt in many industries in recent years, including finance, manufacturing, education, healthcare, and transportation, to name a few.

But there are many more dimensions to the value of AI, and accelerating basic scientific research, boosting social productivity, and improving human living space are just a few of them, so we might as well boost our own imaginations first.

What kind of change will AI bring to humanity, let's wait and see!