This article is from WeChat official account:qubit (ID: QbitAI), author: Xiao Xiao, Guhao Nan, fish, sheep, head Figure from: vision China

The answers to various questions about “artificial intelligence” you want to know are all here.

Professor Liu Yunhao from the School of Software of Tsinghua University spent dozens of hours to integrate the 2,000 questions about artificial intelligence from the students into 92 questions, and then wrote a 30,000-word long article to give answers to these questions.

Whether it is GPT-3, brain-computer interface, or the impact of Sino-US relations on the computer industry, Professor Liu Yunhao has given his own views on these issues.

As the earliest scholar in China to define “Internet of Things”, Professor Liu Yunhao is also a national outstanding young scholar and a Changjiang scholar.

In addition, Professor Liu Yunhao is also the honorary chairman of the ACM China Council and the chairman of the ACM China Turing Conference. He also has considerable research in the field of artificial intelligence.

This interpretation of his can be described as rave reviews:

It is helpful for students in various fields to understand artificial intelligence, very professional and in-depth:

Even people in non-professional fields can understand the current development of “artificial intelligence”:

So for this rare “science popularization”, we extracted and edited the core part of the content, and there is no lack of topics that are most concerned by the current scientific and technological circles. I hope that you who care about artificial intelligence, care about cutting-edge technology, and care about the transformation of the technology industry helpful.

How to develop computer-related industries under pressure from the United States

Q: What is the current status of China’s software industry? How to develop the computer industry under pressure from the United States?

Liu Yunhao: In terms of application software, especially consumer-oriented application software, our country is at a world-class level. Some mobile applications are (such as TikTok) is already at the leading level in the world.

But in industrial software (such as electronic design automation EDA software) and basic software (such as Windows-like operating systems), there is still a big gap between China and the world’s first-class level.

In these areas, international cooperation is very important, because not every technology can be mastered in a short time, but this does not mean that we have to give up independent research and development.

In the face of changes in the world situation, procurement cannot solve all problems, and even legitimate business activities will face interference from the outside. Make friends extensively, embrace the concept of openness, strengthen independent research and development, and gradually reach the world’s leading level. This requires the continuous efforts of several generations, including you and me.

Q: How do you treat Intel’s database leaks? Will it have a good impact on related industries in China?

Liu Yunhao: The leaked information is only part of Intel’s customer information, which is of little value to my country’s chip design.

In the field of chip design, China has made great progress. Whether it is embedded microprocessors, mid-to-high-end system-on-chips, and server processors, a number of domestic companies have developed and expanded. In some areas, such as neural network processors (NPU), the design capabilities of Chinese enterprises are already at the top of the world.

But design ability alone is far from enough. In the high-end chip field, China lacks more refined manufacturing process

Q: Artificial intelligence can play Go, write poetry, etc., even better than humans. Does this mean “machines can think”?

Liu Yunhao: It depends on how we define thinking. If you can talk with people or complete some basic tasks of interacting with people, it is called thinking ability, then many voice assistants, AlphaGo and other AIs already have particularly elementary thinking skills.

And more advanced thinking, people are still working hard. But being able to think is not the same as being conscious. There is still a big gap between AI and human intelligence under current technological conditions.

Q: How far is the current AI from autonomous reasoning research science? Is there a limit to human cognition, and is there a limit to machine cognition?

Liu Yunhao: Scientific research is not just a matter of reasoning. A large number of disciplines rely heavily on experiments. Therefore, it is absolutely not enough to rely on AI to study science.

At this stage, there have been some good results in machine proof. You can learn about the work of Academician Wu Wenjun. On the other hand, the current AI is the product of human invention, and it cannot break through the human cognition of the natural world.

Academician Wu Wenjun