On November 1, at the 2024 Sichuan Netcom "Digital Intelligence Pilot" 14th Sichuan Netcom "Entering the Frontier of New Productivity" series of activities and the 10th Biomedical Big Data·Intelligent Technology Conference, academician of the Chinese Academy of Sciences, Chen Runsheng, a researcher at the Institute of Biophysics, Chinese Academy of Sciences, was interviewed by a reporter from Daily Economic News.
In the interview, Chen Runsheng said that large-scale artificial intelligence models are still in their early stages, including their application in the biomedical industry, and there is still a long way to go. It can be said that the application of large artificial intelligence models in the biomedical industry has just begun.
In the future, "the application and intervention of large artificial intelligence models to the entire medical system will be comprehensive, and the application of artificial intelligence will be realized before, during and after treatment. Artificial intelligence will not only greatly improve the efficiency of medical treatment, but also It will fundamentally change the entire medical system, turning it into a medical supervision covering all people and all stages, changing the entire medical paradigm,” he said.
Recently, the 14th Sichuan Netcom "Digital Intelligence Navigation" 2024 Sichuan Netcom "Entering the Frontier of New Productivity" series of activities and the 10th Biomedical Big Data·Intelligent Technology Conference were held in Chengdu. At the meeting, representatives from domestic and foreign medical experts, smart medical companies and other parties discussed the high-quality development of new health digital productivity.
As one of the earliest scientific researchers engaged in theoretical biology and bioinformatics research in my country, Chen Runsheng said that in general, large artificial intelligence models are still in their early stages, including their application in the biomedical industry, and there is still a long way to go. Gotta go. It can be said that the application of large artificial intelligence models in the biomedical industry has just begun.
"These early data applications, including medical record management, recording of basic registration information, and management of electronic medical records, all used big data to automate the process. With the help of this data, we can analyze the regular parts and then solve the problem. More practical issues. Although still in its early days, big data has already made substantial contributions to the biomedical industry," he said.
Taking the early development of drugs as an example, past experience is that it takes 10 years and US$1 billion to develop a new drug. But with the help of big data and artificial intelligence, the types of compounds that need to be screened may have changed from tens of thousands to hundreds or even dozens, and the search scope has become 1% of the original. The efficiency of the early development of new drugs has been greatly improved. These are the application practices of big data and artificial intelligence in the field of biomedicine.
In Chen Runsheng’s view, all large industry models rely on computing power and data.
"First of all, the key to whether a large industry model can be made is how much industry data the builder has mastered, so data is the key. But with data, two problems need to be solved, one is the standardization of data, and the other is the integration of data ." he said. The so-called standardization of data refers to the universality and mutual recognition of data. If the standards for data generated by various institutions or platforms are not unified, the basis for application will be lost. The integration of data lies in breaking through the limitations of single data. If data sharing cannot be achieved, the role and significance of large models will decrease.
To solve the problems of data standardization and integration, there must be a leading body. Chen Runsheng believes that taking the United States as an example, the main body to solve data standardization may be Open AI, while taking medical industry data as an example, relevant departments such as health care may need to take the lead in solving the standard specification issue of data sources. In addition to solving the problem of data standardization, data integration also requires such an institutional department to take the lead.
In addition, for medical institutions, building their own large pharmaceutical models is still a cost item. For a large number of hospitals that are struggling with profitability issues, how to build and use big data and big models is a matter of cost and benefit output. In this regard, Chen Runsheng said, "The improvement of hospital awareness and the intervention of management departments will gradually solve this problem. Because the use of big data is inevitable for development, if you do not take this step, you will be gradually eliminated. This is not a matter of whether to do it or not. , The question of when to do it is a trend that must be adapted to. Whoever does it first will have the advantage, and whoever does it first will benefit more."
Chen Runsheng said: "The application and intervention of the large artificial intelligence model to the entire medical system is comprehensive. The application of artificial intelligence will be realized before, during and after treatment. Artificial intelligence will not only greatly improve the efficiency of medical treatment, but also It will fundamentally change the entire medical system, turning it into a medical supervision covering all people and all stages, changing the entire medical paradigm. "