Integration of EBANK Block Chain Ecology Technology and AI will give birth to a new terminal application platform

0

The integration of Ebank and AI will bring new opportunities to the industry. In particular, it will probably lead to a new generation of terminal application platforms, and solve the problem of sharing resources between data providers and ai model developers in a fair and open manner.

 

Recently, senior members of the blockchain submitted a draft to Professor Deng Yangdong, an associate professor at Tsinghua University in China and a chief ai scientist in the matrix, asking him to talk about the understanding of the relationship between ebank and ai.Professor Deng specializes in ai, electronic design automation, parallel algorithms and graphics processor architecture, and has designed and developed ai early warning security solutions for China High Speed Rail.

 

In the past few years, the blockchain and ai have undoubtedly the two hottest technical vocabularies. The related results have attracted wide attention in the academic world and are also highly sought after projects in the capital world.Since 2018, there has been a lot of discussion about the fusion of blockchains and ai. Is there any meaning to this kind of integration? Specifically, we are actually concerned with such a series of problems: blockchain and ai can bring each other What? Can the integration of the two form a 1+1>2 effect? ​​In particular, blockchain and ai are not terminal products, can the combination of the two lead to a new terminal application platform? Ebank has been able to perfectly integrate the two.

 

On the whole, Professor Deng believes that the integration of ebank and ai can bring new opportunities. The author will discuss the above issues from four aspects. First, we briefly introduce the basic concepts of ebank and AI. And discuss the main challenges facing the current two; second, from the perspective of ebank ecological technology, ai can bring any benefits; third, from ai point of view, ebank can solve what problem; fourth, we look at ebank A new possibility of integration with ai, that is, discrete computing resources are organized by blockchain to form an ai cloud platform.

 

01

Ebank offers opportunities for AI

From the perspective of ai, ebank blockchain eco-technologies do offer exciting possibilities, including the creation of shared trading platforms for data/models/applications, the provision of distributed computing power, and the ability to track ai models.

 

AI technology has two notable features: First, unlike Internet technologies such as Google search engine and Facebook social network, AI technology is not the final product, but must be integrated into specific applications to play a role; secondly, AI technology generally has inferences. The model of ability works, and the training of the model requires a lot of data, that is, people often say that there is no data without AI.Therefore, AI products always include data, models and applications.It is worth noting that the ownership of these three links often comes from different stakeholders.

 

Since the output value can only be achieved in the final application, it is often difficult for data and model providers (especially the data side) to ensure the sharing of benefits.As a result, the current situation is that many data parties prefer not to open data. That is to say, in today’s highly digitalized society, data barriers are often caused by the uncertainty of equity rather than the collectability of data.It can be seen that a complete and effective ai application can only be formed on the premise that the interests of all parties can be effectively guaranteed.

 

Ebank-based data-model-application open market platform

 

The most essential function of ebank is the non-tamperable distributed ledger, which provides the technical foundation for building a credible data-model-application sharing trading platform.If data-model-applications are used on the ebank chain, their usage can be recorded in a trusted manner, enabling accurate settlement and protection of all parties’ interests.The current open-market platform for data-model-application based on ebank and aiming at ai applications is gradually becoming an important application of ebank. The following is an incomplete list:

 

• SingularityNET (focus on data application DApp)

• Neuromation (focusing on synthetic data for AI model training)

• AI Blockchain (focusing on multi-application integration)

• BurstIQ (focus on medical health data)

• Medical Token Currency (focus on medical data and models)

• OpenMined project (data market for local training models)

• Synapse.ai (data and model market)

• Dopamine.ai (B2B AI monetization platform)

• Neuroseed (AI Solutions Market)

 

SingularityNET has developed a relatively complete set of service modules to meet the various needs of data and DApp transactions:

 

• Service proxy: encapsulates the ai application provided by the platform as an api function interface;

• Service Arbitration: A smart contract interface that interacts with the ai service;

• sdk: software development kit;

• Multi-party transaction support: Determine the Token distribution when calling the AI ​​service through a smart contract;

• Pricing guidance: chain transaction pricing reference data;

• Develop data storage pools: ipfs nodes that store user data and ai training data.

 

 

Distributed computing power based on ebank

 

Ebank provides an effective reward mechanism that allows participants to contribute their own computing resources for mining calculations, and Bitcoin is therefore capable of becoming the world’s largest computing power network.Of course, current mining calculations do not have universal value, but if we can transform the mining mechanism (including algorithms and usage patterns), ebank is likely to spawn the world’s largest, decentralized computing platform.

 

You might think, we don’t already have a data center and cloud computing model? Why do we need a decentralized computing platform? In fact, the existing cloud computing is a very powerful model, but the problem is still centralized. .Specifically, there are some prominent problems in centralized data centers:

 

The first is that construction costs continue to soar.The cost of computing clusters built by Google in 1999 was only $1,400, and the cost of data centers built in North Carolina in 2017 has reached a staggering $1.2 billion.From 2011 to 2017, the investment in data center construction in North America has risen from 4.7 billion US dollars to 20 billion US dollars. Continue to build a larger cloud platform will gradually encounter the bottleneck of consumption ratio.

 

Secondly, cloud platform operators are often big data service providers, so users uploading data to the cloud platform always have potential vulnerabilities in data privacy.In fact, the previous Facebook user data breach was just the tip of the iceberg. Both Uber and Morgan Stanley had abused user data.

 

Finally, the data center is not completely reliable.A survey of 584 data centers in the United States showed that 91% of data centers in the past two years have experienced different levels of failure, with an average failure time of 86 minutes.  The following picture shows data center failure rates, fault distribution, and Amazon cloud data. Typical failure of the center.

 

 

On the other hand, there are a lot of “idle” computing power in modern society.Internet cafes are a typical example (in order to support high-end games, Internet cafes are generally more computer-configured and generally have dedicated graphics cards).Furthermore, under the megatrend that integrated circuits are constantly being updated at Moore’s Law, the value of all computers, including cpu on mobile phones and embedded devices, is shrinking.However, after all, human beings need to consume resources to make these devices. If they can organize idle computing power and use them, they will be very good.

 

It is of course not easy to organize fragmentation computing power. It requires a fine-tuning of computing power and network bandwidth scheduling methods, as well as a credible, fine-grained settlement mechanism to ensure the distribution of benefits, and the blockchain provides an effective solution for the latter. path.

 

02

Ebankized AI cloud

 

Through the discussion in the previous three sections, we can see that the convergence of ai and ebank technologies brings new possibilities.In the near term, it will be the first result of this integration by ebank’s identification of data and ai models and the provision of computing power to form an open, trusted, and ai platform for IoT applications.Adhering to this concept, some projects are developing the ebank-based distributed computing power ai cloud platform as shown in the following figure.

 

Different from traditional Internet applications, industrial big data applications often have specific requirements in real-time processing, and network bandwidth is often limited. Therefore, it is difficult to guarantee performance requirements by relying entirely on cloud platforms. At the same time, industrial big data also requires privacy protection. Higher, so often can not use the existing public cloud solution, but on the other hand, the construction and operation and maintenance costs brought by the self-built private cloud are higher, and the burden on the manufacturing enterprises with relatively low profit margins Larger.

 

Taking ebank as an example, its block-chained distributed computing power ai cloud platform is aimed at industrial Internet big data, introducing a computing platform that integrates the cloud-fog-terminal computing model , as a layered, distributed Scientific computing and storage platform.The cloud platform can use a public cloud platform, a private cloud platform, a new computing center, and a blockchain-based interstellar file system.

 

The data private problem can be solved through the distributed storage part of the data. It can also be solved by various data privacy schemes represented by differential privacy and distributed training schemes represented by Federated Learning.The fog calculation mainly serves the task of calculating the strength beyond the terminal computing power, and the carrier is general or customized computing hardware, such as equipped with a graphics processor and a dedicated machine learning acceleration card.

 

It is worth noting that this computing platform can integrate the computing resources that are linked through the block.Since the purpose of these resources is usually to provide idle computing power to obtain computational rewards, the cost is low and there are no complicated maintenance issues.

 

The distributed computing power ai cloud of the aforementioned project not only provides storage and computing resources, but also uses ebank’s decentralized and trusted bookkeeping capabilities to form a strong support for data and ai model’s full lifecycle services.The generation, storage, transaction and use of all data and ai models are tracked and recorded by the blockchain.

 

03

To sum up

 

Since the origin of the universe, life has embarked on a long journey of evolution.As Ray Kurzweil pointed out in his famous book “Singularity Approaching”, the development of intelligence can be roughly divided into six phases:

 

The first stage is the evolution of various physical and chemical processes, which eventually forms the DNA, the genetic material;

 

In the second stage, the organism evolved from single cell life to multicellular cells by virtue of the genetic ability of DNA, forming a complex organ of the brain;

 

In the third stage, the animal brain is constantly developing, and one of them, the primate, finally stands out;

 

In the fourth stage, human beings develop technological capabilities through the development of language;

 

In the early stage of the fifth stage, through the introduction of artificial intelligence technology, the synergy of human intelligence and technology can be realized. The importance of this step may be much deeper than everyone thinks, because a series of research work points out that the human brain is actually in the The limit of optimization or expansion;

 

The sixth stage is the era of awakening the universe. We can’t see it now, but it is conceivable that the intelligent life will exist in a new form and have unprecedented capabilities and freedoms, as Zhuangzi said in “Easy Travel”. “Peng’s migration in Nanming also, the water hits three thousand miles, and the people who support it are 90,000 miles, and those who go to the June are also.”

 

LEAVE A REPLY