EpiK Protocol: From Recognition to Cognition, Helping AI Broaden Its Horizons
A few decades ago, an IBM computer called Deep Blue beat the world chess champion. In recent years, the Google computer called AlphaGo has successfully defeated the Go world champion and runners-up, in addition to the popular machine translation, autonomous driving, intelligent robots and so on at moment, artificial intelligence is driving the wheel of history to the age of full-scale intelligence, just like steam engines in the Age of Steam, electric generator in the Age of Electricity, and computers and the Internet in the Age of Information.
On the long way of the development of artificial intelligence, it will go through several important stages, from “computation” to “perception”, and then upgrade to “cognition” and “consciousness”. Although artificial intelligence has developed to a certain extent at present, it is still not comprehensive, as Zhang Bo, an academician of the Chinese Academy of Sciences and director of the Institute of Artificial Intelligence at Tsinghua University, said that if the Alps’s images were added with the noise, the computer would automatically identify it as a dog for reasons that could not be explained, which suggests that the current AI suffers the limitations of cognition.
AI faces a new challenge, in urgent need to evolve from perceptual intelligence to cognitive intelligence, which has forced us to re-examine artificial intelligence and find a solution to enable AI cognition.
The Past, the Present and the Future of Artificial Intelligence
Artificial intelligence helps humans do all kinds of things by endowing machines with human intelligence. Since it is artificial intelligence, it should be like a person at first, capable of looking, listening and thinking, then making a decision, and forming the own way of thinking. And in each stage of AI development, these bright colors exist.
The first stage, the stage of computation.
In 1956, the scientist McCarthy, Minsky et.al held a meeting at Dartmouth College of US to discuss “how to simulate human intelligence with machine” and first proposed the concept of “Artificial intelligence, short for AI”, marking the birth of the discipline of artificial intelligence.
With the introduction of the concept of artificial intelligence, some research achievements have appeared, such as machine theorem proving and checker playing program, where artificial intelligence is just a function of computer such as representation, storage and human-computer interaction. During this time, it has become a landmark event that the IBM computer called Deep Blue beat the world chess champion Kasparov.
The second stage, the stage of perception.
Artificial intelligence has entered the stage of perception, which is characterized by being based on ubiquitous perceptual data and deep learning, with image classification, speech recognition, human-computer games and driverless technology as its representative technologies, such as the Google computer called AlphaGo which is the important product in this stage.
The so-called ubiquitous perceptual data consists of three main core contents, of which the first one is machine perception, the second one is computer vision, the third one is natural language processing. These perceptual data help machines evolve in terms of speech and vision, and realize physical perception in terms of listening and viewing.
At this stage, although the machine could better identify things through deep learning and other technology, it still stayed at the recognition level. The situation mentioned by academician Zhang Bo in the previous part of the text often occurs.
At the Scientific Frontier Plenary Session of the World Artificial Intelligence Conference 2020, Guo Yike, the tenured professor at Imperial College London and academician at the Royal Academy of Engineering, believed that artificial intelligence today is just about being able to recognize something, it is the field of the next generation to really know something.
The third stage, the stage of cognition:
Artificial intelligence enters the stage of cognition, that is, machines do the right thing like human beings. How can it work? It can be achieved by making the machine the ability of understanding, thinking, learning, reasoning and making decisions like human beings.
At this stage, machines can become important assistants of humans, such as home intelligent doctor robots, home intelligent legal robots, and so on; not only do mechanical manual work simply and repeatedly, but also replace many complex mental work of human beings, releasing more intelligent human wisdom and inspiration.
Knowledge is a key factor in enabling machines to have cognitive intelligence. Unlike the physical perception function of perceptual data, knowledge is the data of exploration of human cognition, such as payment, searching, creation, language, and so on.
But the machine can’t read knowledge, and the Mapping Knowledge Domains is the solution of structured knowledge data. Mapping Knowledge Domains is a large-scale semantic network that uses a graph model to describe the relationship between knowledge and all things in the world, and a semantic network is an important communication form between machines. For this reason, Mapping Knowledge Domains is an important infrastructure to promote artificial intelligence in the stage of cognition. As the Mapping Knowledge Domains becomes denser the knowledge cognition acquired by machines becomes more, the learning and reasoning ability is stronger, and the machine becomes more intelligent.
The fourth stage, the stage of consciousness
With the improvement of artificial intelligence cognitive ability, through the deepening of self-learning of cognitive ability, the memory mechanism is formed, and with the development of (unconsciousness) consciousness processor, artificial intelligence will realize higher-level intelligentization of human society, and then realize comprehensive intelligence.
Move Towards Cognitive Intelligence EpiK Mapping Knowledge Domains help AI broaden its horizons
As an important infrastructure of artificial intelligence, Mapping Knowledge Domains play an important role in promoting artificial intelligence to move towards cognitive intelligence. It is a huge project to transform accumulated knowledge of human beings into Mapping Knowledge Domains, but it is also the only way to promote the evolution of weak artificial intelligence to strong artificial intelligence. However, the construction of Mapping Knowledge Domains also faces many challenges:
First of all, because human beings have a huge amount of knowledge data and need a large amount of human and material resources, it is an urgent problem to organize these human and material resources.
Secondly, although many Internet giants have joined in the construction of Mapping Knowledge Domains, each of them has their own way, and can not achieve interconnection, resulting in repeated construction, and the lack of effective incentive measures, which is not conducive to the construction of Mapping Knowledge Domains.
Thirdly, the current centralization storage faces the risk of malicious attack, which also puts the risk of Mapping Knowledge Domains being tampered with, and cognitive intelligence needs more secure and credible Mapping Knowledge Domains.
To this end, the establishment of a universal and secure Mapping Knowledge domain collaboration platform has become an important choice for the future development of cognitive intelligence.
EpiK Protocol is a secure and reliable Mapping Knowledge Domains collaboration platform that is decentralized, jointly developed and shared, blockchain solution is used to construct new cognitive intelligent infrastructure, which has four core capabilities: Trust Storage (Filecoin), Trust Incentives (Token Economy), Trust Governance (DAO) and Trust Finance (DeFi).
Among them, Trust Storage brings Filecoin ecological resources in, Trust Incentives proposes that the bigger the contribution, the higher the Trust Incentives, Trust Governance develops by using distributed autonomous organization, and Trust Finance combines the opportunities of market frontier DeFi. EpiK is trying to broaden AI’s vision and empower machines the cognitive intelligence by building a super-large Mapping Knowledge Domains while taking equity into account and realizing joint construction and benefit-sharing.
· EpiK APP Download
· EpiK Twitter and Telegram Channel (For the latest news)
· EpiK Telegram Community
· EpiK Reddit, Facebook, and Discord
· EpiK Medium (For the latest articles)
·EpiK GitHub (For the full set of code)
