In this article, the author introduces the possibility of integration of the two popular technologies of artificial intelligence and blockchain, and what are the standard definitions after the integration of the two, what challenges will be encountered, and what benefits will be brought. The following is a translation
It is undeniable that artificial intelligence and blockchain have promoted innovation and have caused fundamental changes in different industries. The technical complexity of these two technologies is different and the business meaning is different, but if the two can be integrated, the entire technology (and human) paradigm may be redefined.
I have already introduced some content about artificial intelligence, so I won't waste any space here (if you want to know more, you can check out my explanation of artificial intelligence and a brief history of artificial intelligence).
But I have never been exposed to blockchains and cryptocurrencies, so I will cover what these are in the first chapter and what they work.
A blockchain is a secure distributed database shared by all parties in a distributed network. Transaction data can be recorded and easily audited. In short, blockchain is a technology that "allows people who don't know each other to trust events together."
The data is stored in a rigid structure called blocks, which are connected to each other by a hash chain. These blocks have a header and a content part, where the header contains metadata and the content part contains data for real transactions. Since each block is interconnected with the previous block, it is very difficult to modify whatever information you want without the network consensus as the number increases.
The network can verify transactions through different mechanisms, but there are only two main mechanisms: "proof-of-work" and "proof-of-stake". Workload proof (Nakamoto, 2008) requires participants (called “minersâ€) to solve complex mathematical problems in order to add a block, which in turn requires a lot of power and hardware capabilities to decode. Proof of equity (Vasin, 2014) attempts to address this energy efficiency issue, thus placing more mining rights on participants with more money.
There are also other mechanisms, such as the Byzantine fault-tolerant algorithm (Castro and Liskov, 2002), Quorum slices (Mazieres, 2016), and various derivatives of proof of equity, which we will not introduce here.
Regarding the characteristics of the blockchain, the last thing to note is that it can be classified according to different network access rights, for example, can anyone browse (unlicensed Vs requires permission), or participate in the formation of consensus (public Vs) private). In the former case, anyone has read and write access, and in the latter case, the pre-determined participant has the right to join the network (of course, in the case of public without permission, as a reward structure for the miners) Row).
The situation should be clear now. The essential strength of this technology is not only subversive technology, but more importantly it is the basic technology aimed at “changing the scope of the intermediaryâ€. Distributed general ledger technology does reduce the cost of verification and networking, which in turn affects the market structure and ultimately leads to new markets. In recent work, IansiTI and Lakhani (2017) also made a wonderful comparison between the blockchain and TCP/CP technologies, showing how the blockchain has slowly experienced TCP/IP and the like. The four phases that previous fundamental technologies have experienced are the single use phase, the localization use phase, the replacement phase, and the change phase. As they explain, the “novelty†of such technologies makes it difficult to understand the solution domain, and its “complexity†requires a larger institutional transformation to foster a more convenient adoption atmosphere.
However, it is also true that the blockchain is transforming the traditional business model and is shifting its value toward the direction of the previous technology stack: if investment applications are more meaningful than investment agreement technology 15 years ago, In the world of blockchain, value will be concentrated in the shared protocol layer, and the profit level at the application layer will be very thin (see Joel Monegro's "fat agreement" theory).
This is a stack of "fat" and "thin" applications (Joel Monegro)
Finally, as the end of this introductory chapter, I would also like to mention that the possibility of a blockchain in reality is not only limited to transactions, but also to the establishment of (smart) contracts triggered by special events and thresholds that can be easily traced and audited. There are also possibilities.
Additional Information: First Token Release (ICO)
A major hype surrounding this new phenomenon is the first token sale (ICO). Even though many people invest in it because the name is reminiscent of the most common (and most valuable) initial public offering (IPO), ICO is nothing more than token sales, and tokens are the smallest functional unit of a particular network.
I hope that ICO experts can forgive my rough definition, but ICO can be said to be a hybrid concept that combines equity allocation, pre-sales/crowdfunding activities, and elements with limited power and application domains.
Introducing new unregulated financing is definitely an interesting innovation, but it also raises some concerns about communities that are not yet ready. I am very happy to hear your feedback, but here I will refine the key points of the ICO assessment:
There is an extra role in the value exchange token, and the only goal of the company to sell tokens is to give the market a bad signal. Tokens are used to build user groups and motivate stakeholders to participate in the ecosystem at the earliest stages. It is not enough to have a good white paper;
Beware of unrestricted token sales;
Beware of token sales without time limits;
Beware of token sales that don't explicitly state (current and future) quantities and token values ​​(which sounds ridiculous, but you might be surprised by ICO's opacity).
II. How will AI change the blockchain?Although the blockchain is extremely powerful, it also has its own limitations. Some of them are technology-related, while others come from an old-fashioned culture inherent in the financial services field, but all of this is affected to some extent by AI:
Electricity consumption: Mining is an extremely difficult task that requires a lot of electricity (and money) to complete. AI has proven to be an effective means of optimizing power consumption, so I think similar results can be achieved in the blockchain. This may lead to a decline in investment in mining hardware;
Scalability: The blockchain is steadily evolving at a rate of 1MB per 10 minutes, with a cumulative total of 85GB. Nakamoto Satoshi (2008) first proposed that "blockchain pruning" (such as deleting unnecessary data about completely consumed transactions) as a possible solution, but AI can introduce new decentralized learning such as federal learning. System, or introduce new data fragmentation technology to make the system more efficient.
Security: Even if the blockchain is almost impossible to attack, the deeper layers and applications of the blockchain are less secure (such as DAO, Mt Gox, Bitfinex, etc.). The incredible progress in machine learning over the past two years has made AI an excellent ally in the blockchain to ensure secure application deployment, especially given the fixed nature of the system architecture;
Privacy: Privacy issues with personal data raise regulatory and strategic concerns about competitive advantage. Homomorphic encryption (operating directly on encrypted data), Enigma projects, or Zerocash projects is definitely a possible solution, but I think this issue is closely related to the previous scalability and security issues, and I think they The same degree of importance;
Efficiency: Deloitte (2016) estimates that the total operating cost of blockchain verification and shared transactions is approximately $600 million per year. An intelligent system may eventually calculate the possibility of a particular node becoming the first node to perform a particular task in real time, from allowing other miners to choose to abandon efforts for that particular transaction, thereby reducing the total cost. In addition, even with some structural constraints, better efficiency and lower energy consumption may also reduce network latency and allow transactions to be faster;
Hardware: Miners (not necessarily companies or individuals) put incredible money into specialized hardware components. Since power consumption has always been a key issue, many solutions have been proposed and more will be introduced in the future. As long as the system becomes more efficient, some of the hardware may be converted (sometimes partially converted) for use by neural networks (the mining giant Bitmain is doing this);
Lack of talent: This is a leap of faith, but again we are trying to automate the science of data itself (which is unsuccessful according to my current perception), and I don't see why we can't create virtual ones that can create new ledgers. Agent (even affecting and maintaining the ledger);
Data Gate: In the future, when all our data is on the blockchain, when the company can buy directly from us, we will need help to access the authorization, track the data usage, and usually need to figure out the speed of the computer. What happened to the personal information. This is the job of a (smart) machine.
III. How the blockchain changes AI
In the previous section, we quickly touched on the impact that AI might end up on the blockchain. Now let's look at how the blockchain might affect the development of machine learning systems. To be more careful, the blockchain can:
Help AI explain itself (and let us believe it): The AI ​​black box has encountered interpretable questions. A clear audit trail not only improves the credibility of the data, but also improves the credibility of the model, and provides a clear path for retrospective machine decision making.
Improve the effectiveness of artificial intelligence: Secure data sharing means more data (and more training data), then there will be better models, better actions, better results... and better New data. In the end, the network effect is the most important thing.
Reduce barriers to market entry: we come step by step. Blockchain technology protects your data. So why can't you store all your data privately, or maybe sell it? You may be. So first, the blockchain will promote the creation of cleaner, more organized personal data. Second, the blockchain will promote new markets: such as the data market (which is easier to implement); such as the model market (which is much more interesting); and even the AI ​​market may eventually appear (see Ben Goertzel trying to use SingularityNET) Things to solve). As a result, simple data sharing and new markets, along with blockchain data validation, will provide smoother integration, reducing the barriers to entry for small businesses and reducing the competitive advantage of technology giants. In our efforts to lower the barriers to entry, we have actually solved two problems, namely, providing more extensive data access and a more efficient data monetization mechanism;
Increase trust in labor: Once some of our tasks are handed over to automated virtual agents for management, a clear audit trail will help the robots trust each other (and help us trust them). With sub-item data and coordinated decision-making, coupled with a robust mechanism to reach a quorum (highly correlated with group robots and multi-agent scenarios), this will eventually increase machine-to-machine interaction (Outlier) Ventures, 2017) and trading. Rob May also expressed a similar concept in his recent email newsletter.
Reducing catastrophic risk: AIs written in DAO with specific smart contracts can only perform those actions, and there are no more (then its mobile space is also limited).
Although the interaction of AI with blockchain technology can bring many benefits, there is still a big problem that plagues me.
AI was born in an open source environment where data is the real moat. But with the democratization of this data (and the open source of software), how can we ensure that AI is successful and growing? What will the new moat be? My only guess at this stage is... talent.
IV. Decentralized smart company
There are many start-ups working on blockchain and cryptocurrencies. However, I am only interested in the intersection (or integration) of AI and blockchain technologies. These companies are obviously not many. Such companies are mainly concentrated in San Francisco and London, but there are also examples in New York, Australia, China and European countries.
The number of such start-ups is really too small, so it is difficult to classify them further. I usually like to try to understand the impact of the underlying model of a group of companies on the industry / application type, but given the small number of data points, it is difficult to conduct such analysis, so I simply follow the following Sorted out:
Decentralized intelligence: TraneAI (training AI in a decentralized manner); Neureal (point-to-point AI supercomputing); SingularityNET (AI market); NeuromaTIon (integrated dataset generation and algorithm training platform); AI Blockchain (multiple application intelligence) ); BursTIQ (healthcare data market); AtMatrix (decentralized robot); OpenMined project (data market for training machine learning locally); Synapse.ai (data and AI market); Dopamine.ai (B2B AI monetization platform) );
Conversational platform: Green Running; Talla (chat bot); doc.ai (quantitative biology and healthcare insights);
Forecasting platform: Augur (collective intelligence); Sharpe Capita (crowd sentiment forecast);
Intellectual property: Loci.io (IP excavation and mining);
Data traceability: KapeIQ (fraud detection for healthcare entities); Data Quarka (facts verification); Proops (data compliance); Signzy (KYC)
Trading: Euklid (bitcoin investment); EthVentures (investment in digital tokens). Other (theoretical) financial applications can be found in Lipton (2017);
Insurance: Mutual.life (P2P insurance), Inari (general insurance);
Other: Social Coin (Citizen Reward System); HealthyTail (Pet Analysis); Crowdz (E-Commerce); DeepSee (Media Platform); ChainMind (Network Security).
EvaluationIt is interesting that many AI-blockchain companies have advisory boards that are larger than the team size. This may be an early sign that the integration has not yet been fully completed, indicating that we don’t know more than we know;
I personally are very excited to see the development of the first type of start-ups (decentralized intelligence), but I also see the huge development of conversational platforms and forecasting platforms and intellectual property. I classify other examples as "miscellaneous" because I don't think that the current stage represents a specific category, but instead a single attempt to match the AI ​​to the blockchain;
It is extremely difficult to evaluate these companies. These websites are often mysterious, so people can't figure out what they do and how they do it (this is because you buy the blockchain precisely because its transparency is a bit different), and this technology needs to be accepted. High-tech education can fully evaluate it. It is a daunting task to open the fog of hype, and the hype is easy to be fooled. But I can show you a concrete example: Have you heard of Magos AI? In my research on the company for this article, I read several articles about the AI-driven blockchain prediction platform company (from Wired, Prnewswire, etc.), which has just completed more than $500,000. ICO, and a great commitment to its delivery. But if you think they should share the ICO materials/information and want to go to the other website, it's weird that their website can't be opened. Of course, sometimes this happens. But I am still not willing, because I read its article on Wired, I would like to know more. I managed to find out who the co-founder was, but I couldn't find his information in Linkedin. Another strange thing. However, some people don't like social activities, especially if you consider that there are no signs of the company's existence three months ago. Then let's take a look at the other team members. There is no information, and I can't find any traceable evidence about his past experience (except that CTO is a theoretical analysis, but I have not found relevant evidence). I try to dig deeper into their skills: I want to find their white papers, blue books, yellow books, or whatever books. But I can only find relevant comments, but I can't find the text. The last two points: I don't think I am a blockchain expert at all, but I have read a lot of things in this area. And I also believe that I am quite familiar with artificial intelligence and industry dynamics. These guys claim to have built five different neural networks that can achieve the same accuracy in different areas than Libratus (or DeepStack), but I have never heard of such a network - it's very strange . Ok, maybe I can write to them and ask for a face to get to know them. Can you know? Their address is the Zurich office of AXA.
After a five-minute investigation, I finally got two keywords: "Magos scam." These guys seem to take away the money and run away. They must have gone somewhere to build that neural network. So please pay close attention.
My point is that exponential technology is very good, it can promote human development, but as its benefits increase, the potential "negative integration" will also grow exponentially. Be alert.
V. ConclusionBlockchain and AI can be said to be two extreme aspects of the technology world: one is to foster centralized intelligence on a closed data platform, and the other is to promote decentralized applications in an open data environment. However, if we can find the right way for these two technologies to work together, the total positive externalities can be amplified in an instant.
Of course, due to the integration of these two technologies, it also has an impact on technology and ethics. For example, how should we edit (or even forget) the data on the blockchain? Is the editable blockchain a solution? After the integration of AI-blockchain, will we become data storage households?
To be honest, I think the only thing that can be done right now is to keep trying.
Lithium Battery Cr2,Film Camera Battery,Industrial 3V Lithium Batteries,Instrument Primary Lithium Battery
Jiangmen Hongli Energy Co.ltd , https://www.honglienergy.com