“The name of the two moves is the same, but the move is very different, a move is the fierce sword of quanzhen jianfa, one is the dangerous family number of Jade jianfa, double sword, power is amazing immediately.” This description from the “Condor Heroes”, said that Yang Guo and xiaolongnu against the golden wheel of the French bridge section, two people with the “jade sword”, still difficult to hostile hand. Yang Guo inadvertently use “quanzhen jianfa”, under the double sword, power increased, kill frequently, unexpectedly won the king of The Golden wheel.

One female and one male, one Yin and one Yang, Yin and Yang are coordinated and powerful. This is the so-called “combination of two swords”. To achieve this, however, there is a high requirement: “If the man and the woman are not lovers, many of the subtleties are hard to hear; The heart can not communicate with each other, then when the sword is too polite friends, elders and younger people are inevitably look at rely on; Such as husband and wife together make, wonderful is wonderful, but it is sentient, yingying shy, half-hearted, worry about gain and loss of all kinds of mood but another layer.” So it is difficult to combine technology.

Today, it is appropriate to compare science and technology to swordsmanship. Technology has its strengths and weaknesses, and different technologies have their flaws. The combination of different technologies, if they can eliminate each other’s flaws, will certainly be more effective. Many examples of “Internet plus” have fully proved this point of view, and the most popular technologies, “AI” and “blockchain”, have also started to cooperate after meeting bottlenecks. Can the two technologies complement and coordinate to achieve upgrading?

Combined with the “Blockchain +AI” industry research report released by the National Internet Financial Security Technical Experts Committee, let’s discuss what kind of sparks AI and blockchain can clash.

“AI” and “blockchain” have their flaws

Although both technologies have been in development for some time, it is still too early to say they have reached a state of sophistication, especially with blockchain, which is still in its infancy.

“AI”

AI is short for Artificial Intelligence. It is a technology that researches and develops for simulating, extending and expanding human intelligence. It has a history of more than 60 years since its birth. In recent years, it has entered the stage of rapid development and can be applied in various fields, but there are still many pain points in the industry.

The growth of ARTIFICIAL intelligence requires huge amounts of data to be trained, and so far companies have collected their own data. Because of the gap in strength and knowledge, the data of the head players is much more abundant than other companies, such as Google, Baidu, Alibaba, Tencent, Microsoft, Apple, Facebook and Amazon. As a result of competition, all the best, so most enterprises are lack of data, less quality data, and the head players also have incomplete data helpless.

In addition, the current ARTIFICIAL intelligence is mostly supervised learning, which requires full labeling of data. However, the labeling in the current industry is mostly outsourced, lacking professionalism and poor quality, and the security and trustworthiness of data are insufficient. And bad data may bring security risks to artificial intelligence. The massive data of artificial intelligence inevitably involves personal privacy information, which poses a great challenge to privacy protection. Artificial intelligence requires a lot of training and takes a long time to train when the authenticity of information needs to be confirmed. Such problems are a serious drag on artificial intelligence.

However, in addition to data problems, computing power and algorithms are also limiting the development of ARTIFICIAL intelligence. On the one hand, hardware costs are high. The training of artificial intelligence in various fields requires a great deal of computation, and the purchase of GPU, FPGA and other hardware resources requires millions of capital, which is unbearable for most ordinary enterprises. On the other hand, due to the lack of elites in this field, it is difficult to update and maintain the algorithm.

“Blockchain”

Blockchain is a new application mode of distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm and other computer technologies. It performs distributed bookkeeping through decentralized, sharing and encryption technologies. It has the obvious characteristics of disintermediation, openness, autonomy, information cannot be tampered with, anonymity and so on.

However, in order to support the realization of these technologies, it needs to pay the price of high power consumption, excess computing power and low efficiency.

Blockchain projects using POW consensus mechanism need to consume a large amount of power resources. To realize the function of distributed data storage, blockchain requires a large number of nodes to run for a long time. The larger the scale, the power consumption will increase proportionally. If you really want to use blockchain in more areas, such consumption is obviously unrealistic.

Behind such large consumption, it is difficult to calculate the low utilization rate of force resources. The computing power of blockchain is not properly utilized, and there is generally surplus or idle state, resulting in great waste, which is exactly the opposite of the insufficient computing power of ARTIFICIAL intelligence.

Each node of the blockchain works too much, resulting in slow efficiency and wasted a lot of cost. Deloitte estimated in 2016 that the total operating cost of blockchain authentication and sharing transactions is around $600 million per year.

Two technologies with their own pain points are joining forces. Will they merge, live in harmony, or repel each other?

Theoretical complementarity

Although BOTH AI and blockchain have their own pain points, they also have many advantages, and in theory, their advantages can make up for each other’s shortcomings.

For ai’s lack of data, blockchain’s massive data can be used. Because blockchain global data can be shared and traceable, data annotation quality is better under such a huge audit work. In addition, due to the distributed storage of blockchain, each node stores complete data information, which also ensures the security of data and improves the trustworthiness of information. Unless all nodes are tampered with, it is difficult to pose a threat to their security.

The anonymity of blockchain is also a good solution to the problem that artificial intelligence may lead to the disclosure of personal privacy. Because blockchain uses asymmetric encryption and authorization technology, while transaction information is open and transparent, account identity information is highly encrypted. So your privacy is protected from prying eyes or being stolen by people with ulterior motives.

Data security and trustworthiness have been guaranteed, artificial intelligence training can also be left a lot of attention. At this point, the method of blockchain distributed data storage can be used to distribute a single model or data on different machines, and model parallel or data parallel training can certainly greatly shorten the training time.

In this way, blockchain is a gift to AI. In return, AI seems to be the solution to blockchain’s woes.

How to reduce blockchain energy consumption through AI? On the one hand, ARTIFICIAL intelligence can replace human mining to complete the task in a more efficient way. It saves manpower and energy waste. On the other hand, the AI learning algorithm can effectively manage the equipment, heat dissipation, cooling and other operations, which can also reduce energy consumption. Applications in this regard have been implemented by Google, Baidu and other companies.

Mining is a complex task, and by using ARTIFICIAL intelligence, it may be possible to calculate the node at which the first task is executed, which may reduce unnecessary exploration by other absenteeism, save more useless work and improve efficiency.

As for the problem of excess computing power of blockchain, artificial intelligence will naturally solve it.

To sum up, blockchain solves the problems of lack of artificial intelligence data, data security, trustworthiness, lack of personal privacy computing power, etc. Artificial intelligence can also make up for the lack of energy consumption and low efficiency of blockchain.

When the two are combined, is it not like the combination of two swords? That said, much of this complementarity is theoretical. The convergence of the two technologies is just beginning.

It actually needs to be run in

For example, Google’s DeepMind Health is developing a blockchain healthcare data audit system that uses “blockchain +AI” technology to enable hospitals, the NHS and patients themselves to track their personal Health data in real time, while protecting patients’ privacy.

AI companies such as Innoplexus have launched blockchain platforms, while Bitmain and Jia Nan Yunzhi are also developing THEIR own AI chips.

Through the confirmation of rights through blockchain, AI identifies the copyright and jointly maintains the copyright, which is also the application of “blockchain +AI”. There are data markets, finance, cloud computing, Internet of things, and so on. But so far, no perfect combination has emerged.

The reason may be that some players are just rubbing their hands in the heat. The main reason is that the industry’s idea of the integration of blockchain and AI is not perfect enough.

The first is data sharing. Maintaining the current status quo allows the two giants to maintain their absolute dominance. AI companies, in particular, have always been ahead of their peers in data, but once AI and blockchain are converged and data is shared, this advantage will no longer exist. The existence of this contradiction can make the marriage of the two less smooth.

And there are risks to merging the two technologies. This is not a simple operation of putting two machines together, but a fusion at the technical level that will be better or worse.

Just as two swords meet well, even if the sword can be fused, it is impossible to be truly united without understanding.

It remains to be explored and studied whether “blockchain +AI” can realize the combination of two strengths, complement each other and upgrade together.

Article/Liu Kuang public account, ID: Liukuang110, this article first leek finance