从AI发展史看“AI + DePin”概念下的IO.NET

In its latest round of financing, the decentralized computing platform based on has reached an estimated value of US$ 100 million. In this month, it was announced that it had completed a round of financing of US$ 10,000 under the leadership of, including the participation of and others. It focused on aggregating resources with machine learning companies and committed to providing services at lower cost and faster delivery time. Since its launch last month, it has grown to more than 100, and has handled more than 100 computing hours for artificial intelligence and machine learning companies. The vision is to build a global. Decentralized computing network builds an ecosystem with machine learning teams, enterprises and powerful resources all over the world. In this ecosystem, computing resources become commercialized, and supply and demand sides will no longer be troubled by lack of resources. In the future, it will also provide access to model stores and advanced reasoning functions such as serverless reasoning, cloud games and pixel streaming. Before introducing the business logic, we must first understand the track of decentralized computing power from two dimensions, one is the development process of computing, and the other is the development process of computing. Latitude is to understand the development process of case computing that used decentralized computing power in the past. We can describe the development track of this computing from several key time points-the early days of machine learning. During this period, machine learning methods mainly focused on relatively simple models, such as decision tree support vector machines. These models have relatively low computational requirements and can run data sets on personal computers or small servers at that time. Feature engineering and model selection are key task time points. From the 1990 s to the early 1990 s, personal computers or small servers with relatively low computing power can meet the demand. Computing hardware dominates computing resources. Second, the rise of deep learning in recent years, the concept of deep learning was reintroduced in this period, which was marked by the research of others. Subsequently, the successful application of deep neural networks, especially convolutional neural networks and cyclic neural networks, marked a breakthrough in this field. At this stage, the demand for computing resources increased significantly, especially when processing large data sets such as images and voices. The victory of Interpoint Competition in this competition is a landmark event in the history of deep learning. It shows the great potential of deep learning in the field of image recognition for the first time. It defeated the world Go champion Li Shishi in, which should be the highlight moment so far. This not only shows the application of deep learning in complex strategy games, but also proves to the world that its ability to solve highly complex problems requires a significant increase in computing power, and it needs more powerful computing resources to train complex deep neural networks. Computing hardware began to become the key hardware of deep learning training, because they are far superior to the era of three large language models in parallel processing. Up to now, with the emergence of years and technical years, large models have begun to dominate the track. These models usually have billions to trillions of parameters, and the demand for computing resources has reached an unprecedented level. Training these models requires a lot of power or more professional and requires a lot of cooling facilities to support them. In addition to special hardware optimized for large-scale machine learning models, a series of computing hardware began to appear. From the exponential growth of the demand for computing power in the past year, the demand for computing power in early machine learning was low, and the demand for computing power was increased in the era of deep learning, and the large-scale model further pushed this demand to the extreme. We witnessed a significant improvement in the number and performance of computing hardware, which was not only reflected in the traditional data center. The expansion of the scale and the improvement of hardware performance are more reflected in the high investment threshold and rich return expectation, which is enough to make the killing between Internet giants public. The initial investment of traditional centralized computing centers requires expensive hardware procurement, such as the construction or lease of their own data centers, the cost of cooling systems and maintenance personnel. In contrast, the decentralized computing platform project has obvious advantages in building costs, which can significantly reduce the initial investment and operating costs. Micro-teams build their own models and create possible decentralized projects. Individuals and enterprises can contribute idle resources to the network, reducing the need for centralized procurement and deployment of high-performance computing resources. Secondly, in terms of operating costs, traditional clusters need to continuously maintain power and cooling costs, while decentralized projects can allocate these costs to each node by using distributed resources, thus reducing them. The operating burden of a single organization is based on the documents, which greatly reduce the operating cost by aggregating the underutilized resources from independent data centers, cryptocurrency miners and other hardware networks. In addition, the economic incentive strategy makes it have great advantages in pricing. Looking back at the history, some decentralized computing projects have indeed achieved remarkable success, attracting a large number of participants and producing important results without economic incentives. For example, this is a project by Sri Lanka. The project initiated by Tamford University aims to simulate the protein folding process through distributed computing and help scientists understand the disease mechanism, especially the diseases related to protein's improper folding, such as Alzheimer's disease, Huntington's disease, etc. During the epidemic, the project gathered a huge amount of computing resources to help study Covid-19. This is an open source software platform that supports various types of volunteers and grid computing projects, including astronomy, medicine, climate science and other fields, and users can contribute idle computing resources to participate. In various scientific research projects, these projects not only prove the feasibility of decentralized computing, but also show the great development potential of decentralized computing. By mobilizing all sectors of society to contribute unused computing resources, the computing power can be significantly enhanced. If the economic model is innovatively added, it can achieve greater cost-effectiveness in the economy. Experience shows that a reasonable incentive mechanism can attract and maintain users' participation, and a mutually beneficial and win-win community environment can be built by introducing incentive models. Step by step to promote the expansion of business scale and promote technological progress in a positive cycle, so it can attract a wide range of participants to contribute computing power through the introduction of incentive mechanism to form a powerful economic model and decentralized computing power potential of decentralized computing network, which not only promotes technological innovation, but also provides value for participants, which can make them stand out from the competition in the field, and has great development potential and market space. A technology cluster refers to complex computing, which connects multiple computers through the network to form a collaborative computing cluster. 比特币今日价格行情网_okx交易所app_永续合约_比特币怎么买卖交易_虚拟币交易所平台

注册有任何问题请添加 微信：MVIP619 拉你进入群

文章转载注明 网址：https://netpsp.com/?id=58433