Look at the frontier
According to a report on the website of the British journal Nature on the 19th, scientists at the University of Washington in the United States used generative artificial intelligence (AI) tools for the first time to help them make new antibodies. According to the research team, AI-designed antibodies may be able to better target some drug targets that are difficult to attack, but these antibodies are still a long way from clinical application.
One of the paper's authors, Nathaniel Bennett, a computational biochemist at the University of Washington in Seattle, and colleagues use artificial intelligence tools released by the team last year to help transform protein design. The tool, called RFdiffusion, enables researchers to design miniaturized proteins that bind tightly to another protein. But these custom proteins bear no resemblance to antibodies, which recognize targets by means of soft rings, which have proven difficult to model with artificial intelligence.
To solve this challenge, the researchers trained and fine-tuned the RFdiffusion neural network using the antibody structures attached to the target identified in thousands of experiments, as well as real-world data from other antibody-like interactions. Using this method, they designed thousands of antibodies. They recognize several bacterial and viral proteins, including those used by coronavirus and influenza viruses to invade cells, as well as a specific region of a cancer drug target.
According to Bennett, AI tools have made it possible to "democratize the ability to design antibodies" by providing a shortcut to high-cost antibody manufacturing. "Ten years later, that's how we designed our antibodies. Charlotte Deane, an immunoinformatician at the University of Oxford in the United Kingdom, said, "This is a very promising study" and represents an important step in the application of AI protein design tools to make new antibodies.
The research team hopes that RFdiffusion can help find drug targets that have proven to be challenging, such as the G protein-coupled receptor, a membrane protein that controls the cell's response to external chemicals. However, the team says that it will take a long time from RFdiffusion to design the antibody to practical application. On the one hand, the binding force of the working antibody to the target is not particularly strong. On the other hand, the sequence of the therapeutic antibody also needs to be modified to be similar to that of the human natural antibody to avoid triggering an immune response.
Image source: Nature website
Technical Value Observation
The upstream of the artificial intelligence industry chain is hardware equipment and data equipment, including chips, sensors, big data, cloud computing services, etc., to provide data services and computing power support for artificial intelligence; The midstream is the core technology of the AI industry, including general technology, algorithm model, and development platform, covering core technologies such as machine perception, natural language understanding, machine learning, graph, and basic open source framework. The downstream is the application products and scenarios of artificial intelligence, involving transportation, medical care, security, finance, home furnishing, manufacturing and other fields.
For the first time, scientists are using generative artificial intelligence (AI) tools to create new antibodies that may better target hard-to-target drugs. Therefore, from the perspective of the artificial intelligence industry chain, the technology is in the downstream application link of the industrial chain.
Macro Market Watch
——The industry is in a stage of rapid development
Compared with the development process of international artificial intelligence, the mainland artificial intelligence industry started late, germinating in 1978. With the rapid development of the Internet and the iterative update of technology, the artificial intelligence industry has gone through a period of exploration and growth in the past nearly 20 years, and is currently in a stage of rapid development. The 13th Five-Year Plan for National Science and Technology Innovation issued by the State Council in 2016 clearly mentions that artificial intelligence will be listed as a major scientific and technological project at the national strategic level. In 2020, the "14th Five-Year Plan" pointed out that artificial intelligence, as an important part of the construction of "new infrastructure", will be the core driving force of a new round of industrial transformation, shouldering the goal of promoting the transformation and upgrading of the trillion-dollar real economy industry.
-- The scale of the artificial intelligence market exceeded 500 billion yuan
Benefiting from the support of national policies and the drive of capital and talents, the mainland's artificial intelligence industry has flourished and has entered the forefront of the world. According to data released by the China Academy of Information and Communications Technology, the scale of the mainland's artificial intelligence industry began to grow rapidly from 2019, with a year-on-year increase of 33.3% in 2021 and a year-on-year increase of 18% in 2022 to 508 billion yuan. According to preliminary statistics, the scale will reach 578.4 billion yuan in 2023, and the growth rate will slow down to 13.9%.
——The AI model is a new paradigm of intelligent computing
The ultra-large-scale intelligent model, or large model for short, is a new AI computing paradigm that has emerged in recent years. Compared with traditional AI models, large models use more data for training, have better generalization, and can be applied to a wider range of downstream tasks. According to application scenarios, AI large models mainly include language large models, visual large models, and multimodal large models. Typical natural language models in the industry include GPT-3, Yuandao, Wudao, and Wenxin. Large visual models have also been widely used in autonomous driving, intelligent security, medical imaging and other fields. The technology of using Wensheng graphs based on multi-modal large models has also developed rapidly, and AI Generated Content (AIGC) has become the next key area of AI development.
-- China's AI applications are accelerating in the financial, telecommunications, manufacturing, and medical industries
At present, with the deepening integration of the digital economy and the real economy, and the transformation of the digital scene of the Internet platform to the metaverse, the overall demand for the total amount and richness of digital content continues to increase. As a new type of content production mode, AIGC has taken the lead in making significant innovations and developments in industries with a high degree of digitalization and rich content demand, such as media, e-commerce, film and television, and entertainment, and its market potential is gradually emerging. At the same time, in the process of promoting the combination of data and reality and accelerating industrial upgrading, AIGC applications in various industries such as finance, healthcare, manufacturing, and industry are also developing rapidly.
According to IDC, the top five industries in terms of application penetration in Chinese's artificial intelligence industry in 2022 are Internet, finance, government, telecommunications, and manufacturing. In addition, the value of AI for autonomous driving and transportation logistics cannot be ignored, according to McKinsey, AI will create 380 billion yuan of economic value for the transportation sector.
-- Forecast of industry development prospects
According to the estimation of the Prospective Industry Research Institute, the AIGC market size in mainland China is expected to reach 17 billion yuan in 2023, and 2023-2025 is the first stage of the growth of the market size of the AIGC industry in mainland China, with a growth rate of about 25%, and the market size will reach about 26 billion yuan in 2025. Starting from 2025, due to the improvement of the industry ecology (especially the completion of the opening of the underlying large model), the vigorous application layer will drive the rapid growth of the industry, with a compound annual growth rate of more than 70%. It is estimated that the scale of the AIGC industry in mainland China will exceed 60 billion yuan in 2027, and starting in 2028, the AIGC industry will extend the complete industrial chain, and continue to expand and deepen the commercialization scene to deeply transform the industry. From 2028 onwards, the scale of the AIGC industry in mainland China will continue to maintain rapid growth, and the market size will exceed one trillion yuan in 2030.
Chinese artificial intelligence technology track heat map
According to the forward-looking industry heat map, the urban clusters strongly related to the key technologies of artificial intelligence are mainly concentrated in South China and Northwest China, and with Guangzhou, Shenzhen and Xi'an as the key development areas, the future layout of artificial intelligence technology and other related technologies The development path of artificial intelligence technology and other related technologies is very likely to be preferentially introduced in South China and Northwest China, which can focus on Tianhe District, Guangzhou City, Guangdong Province, Futian District of Shenzhen City, Guangdong Province and Yanta District of Xi'an City, Shaanxi Province are located in AI-related enterprises, as well as the potential space for the investment environment and supply market for the development of the AI industry.
Prospective Economist APP Information Group
For more research and analysis of this industry, please refer to the "Chinese Artificial Intelligence Industry Development Prospect Forecast and Investment Strategic Planning Analysis Report" of the Prospective Industry Research Institute.
At the same time, the Prospective Industry Research Institute also provides solutions such as industrial big data, industrial research reports, industrial planning, park planning, industrial investment, industrial mapping, smart investment promotion system, industry status certification, IPO consulting/fundraising feasibility study, IPO working paper consulting, etc. To quote the content of this article in any public information disclosure such as prospectus and annual report, formal authorization from Qianzhan Industry Research Institute is required.
More in-depth industry analysis is available in the [Prospective Economist APP], and you can also communicate and interact with 500+ economists/senior industry researchers. More enterprise data, enterprise information, and enterprise development are all in the [Qichamao APP], the most cost-effective and most comprehensive enterprise query platform.