Over-reliance on artificial intelligence and deep learning is not a good thing

The excessive promotion of deep learning using artificial intelligence algorithms has and is causing China huge economic and political losses that are difficult to recover. This is by no means an exaggeration!

Because China still does not have its own general artificial intelligence algorithm. At present, the deep learning artificial intelligence algorithm used throughout China has serious defects. Even the inventor Professor Hinton announced in 2017 that he would abandon it and start over again because of the NP and black box problems that he could not overcome.

Originally, from an academic point of view, deep learning has a position as a representative of a generation of algorithms and has made certain technical contributions in the history of the development of artificial intelligence because it is more advanced than previous artificial intelligence algorithms. In simple application scenarios such as image recognition, face recognition, and voice recognition, deep learning still has certain application effects, but it is far from the magical effects and wide applications that are currently advertised. On the contrary, there are unexplainable problems in deep learning. As experts say, using deep learning, the machine sometimes recognizes the mountain it has learned as a dog. What is more dangerous is that because deep learning has a black box, it is unknown when errors will occur, and the consequences are unpredictable. Therefore, the application field of deep learning algorithms is limited.

With such serious problems with deep learning, international monopoly companies led by Google in the United States saw that deep learning requires infinite computing power and there is a great business opportunity to sell their large-capacity servers, so they began to copy deep learning. The most outstanding commercial copy was the boast that the robot AlphaGo defeated the best human Go player. In fact, only a small amount of deep learning was used in the robot, and a large number of Go rule libraries were still in effect. However, for commercial interests, the monopoly companies deliberately concealed the truth and promoted Shendu Learning to the public, which made Shendu Learning famous for a while.

At the same time, these monopoly companies invested heavily in developing some deep learning open source software for application scenarios, which can be applied quickly. They also used various methods and channels to train a large number of deep learning application technicians. As a result, in a very short period of time, deep learning technicians occupied artificial intelligence positions in various departments and units in China, and promoted the use of deep learning with high salaries.

Therefore, a strange phenomenon has emerged: China, which does not have its own general AI algorithm, is being called upon to vigorously develop AI that has international strategic competition significance, and is actively responding to the call and investing heavily in promoting foreign deep learning. Therefore, why is AI never on the list of high-tech bottlenecks that the United States is blocking China? Because China does not have an AI bottleneck, so why does it need to be blocked?

China is using deep learning, which is promoted by Google and other foreign countries, but we have never seen government departments and research institutions propose to vigorously develop general artificial intelligence algorithms like they have focused on developing operating systems and high-end integrated circuits. I don't know if they don't know about it, or if they know about it but have no choice but to do so due to limited technology. It would be best if it has been included in the plan and is secretly organizing forces for research and development.

In recent years, the country has repeatedly raised artificial intelligence to the national development strategy and included it in the key development areas. Naturally, all departments, localities, and units regard the development of artificial intelligence as a political task. Therefore, all projects related to the development and promotion of artificial intelligence applications, such as product research and development, construction of research units, development zones, industrial parks, etc., are scrambling to approve and invest heavily. Since there is no authoritative organization to promptly explain the true status of artificial intelligence in my country, and only experts who understand deep learning have the final say, a wrong understanding has been formed in society, and when artificial intelligence is mentioned, it is considered deep learning. Therefore, there is a reckless development of artificial intelligence without knowing that deep learning has serious defects, and not thinking about how much money will be sent to monopoly companies due to the large-scale use of backward deep learning algorithms. It is even more unimaginative that a large number of projects implemented around backward deep learning will cause huge losses when they fail to achieve their goals.

In fact, there are organizations and individuals in China who are developing general artificial intelligence algorithms. Two years ago, Dr. Gu Zecang, a director of the China Embedded System Industry Alliance and chief scientist of Tianjin Apollo Technology Co., Ltd., invented the self-disciplined learning SDL general artificial intelligence algorithm to address the shortcomings of deep learning, but it was not recognized or supported after it was announced. Helplessly, Dr. Gu applied the SDL algorithm he invented to autonomous driving, which can reflect the most complex and highest level of artificial intelligence, and reached the world's leading level that exceeds all deep learning teams. Dr. Gu is now raising funds to make SDL modules, and based on the success, he will raise funds again to make the world's first integrated circuit for artificial intelligence algorithms.

We hereby call on relevant departments and society to promptly care about and support Dr. Gu's plan to develop China's own SDL general artificial intelligence algorithm module and integrated circuit. The successful development and industrialization of SDL algorithm modules and integrated circuits will help to recover some of the losses caused by the excessive promotion of deep learning, and will effectively promote my country's voice in the international competition for the development of artificial intelligence.