smart-凯发网

media contact:

ji bin

jibin@cutech.edu.cn

organized by china university science & technology magazine and hosted by intel (china) limited and aliyun computing company limited, china pharmaceutical thinking and sharing conference (the second session) was held in nanjing. with the theme of “smart pharma future”, the conference focused on three major topics: digital transformation and data governance of pharmaceutical enterprises, artificial intelligence (ai)-enabled drug r&d and clinical trials, and the development trend and investment in the pharmaceutical industry.

mr. ji bin, vice president of china university science & technology magazine, ms. liang yali, vice president of marketing group and general manager of cloud & industry solutions department of intel corporation, and mr. huo jia, vice president of alicloud intelligence and general manager of industry solutions sales department, attended the conference and delivered speeches. prof. chen ligong, vice president of tsinghua university school of pharmacy, zhang liang, senior technical director of alicloud intelligence, wu yiming, director of healthcare and life science technology and yang wei, architect of intel, and li zhaorong, general manager of alicloud intelligence digital healthcare products, delivered keynote speeches.

in his speech, jibin said that ai is promoting rather than changing the rules of the pharmaceutical industry, and introduced how china university science and technology magazine can help the application of ai in the field of medicine from the main theme of the conference, how the magazine promotes the development of university science and technology, and how it relies on the existing platform to support the development of the conference in three dimensions. he said that the positioning of the magazine is to promote the development of science and technology in national universities, digital transformation of the authoritative media, the magazine's “four magazines and two networks” (china university science and technology magazine, china science and technology papers magazine, china education network magazine, china science and technology papers online fine papers journal, china science and technology papers online, china academic conferences online) media matrix are respectively the main theme of the conference, and how to rely on the existing platforms to support the development of the conference. the media matrix of “china university science and technology” (china science and technology paper magazine, china education network magazine, china science and technology paper online journal of excellence, china science and technology paper online, and china academic conferences online) promotes the development of science and technology and digital transformation of colleges and universities from the basic support of digitization, the frontiers of natural sciences, and science and technology management of universities. among them, the main task of the market work of “science and technology in chinese universities” is to screen information and promote exchanges around the construction of innovation system, intellectual property management, and transformation of industry, academia, and research, to help scientific research forces to better understand the market demand, to help enterprises to find better scientific research resources, and to promote the capital to find projects that can be quickly realized as well as the rapid realization of the projects; china science and technology papers online platform, as a platform relied on by the conference as the platform on which this conference is based, china science and technology paper online platform has 360,000 registered users and 100,000 accepted papers, and its development focuses on promoting the development of science and technology in the relevant fields in the direction of biomedicine and ai, which are urgently needed by the state to promote scientific research.

in her speech, ms. liang said that over the past three years, a sudden pandemic has brought challenges to many industries, but there are also many industries that have risen to the top. the life sciences industry is one of them. pharmaceutical r&d is at the inflection point of digital transformation and change. intel, as an innovator in the field of basic technology r&d, knows that qualitative changes are driven by the accumulation of quantitative changes, and the continuous maturity and evolution of the ecosystem is one of the most important factors in realizing qualitative changes. today, we bring together colleagues from upstream and downstream of the pharmaceutical industry chain in the hope that through this cross-border communication, we can promote the combination of technology with industry, academia, research and capital, and jointly promote the development of the pharmaceutical industry.

based on his years of experience in providing it solutions for various industries, huo jia pointed out in his speech that the biomedical industry is becoming more and more dependent on computing power. as a cloud computing company, aliyun hopes to make computing power more affordable and ai more popular,” he said. “computing industry” may become a new paradigm for innovation, and we hope to explore more possibilities in the pharmaceutical industry together with the participants through aliyun's self-developed cloud computing platform, and strive to create more health and happiness for this society.

prof. chen ligong gave an in-depth analysis on the topic of “drug target discovery based on slc transporter proteins”. he said that data, computing and ai have brought about profound changes in all industries, affecting every aspect of our lives, and their role in drug discovery and development will also become increasingly important. while displaying the latest achievements of related scientific research projects, he emphasized that modern research cannot be separated from analysis and calculation, and cannot be separated from the support of arithmetic power and algorithms, and all kinds of experiments are closely related to the analysis of data, and the application of ai is also the same, and the essence of it is also data mining, and data mining cannot be separated from the arithmetic power of the brain, but the chip needs to be matched with excellent algorithms, in order to be the most important part of scientific research. but the chip also needs to be accompanied by excellent algorithms in order to become the “strongest brain” in scientific research and innovation.

mr. zhang liang delivered a keynote speech on “data intelligence, continuously helping pharmaceutical enterprises” from the aspects of “digital intelligence technology helping pharmaceutical enterprises' development” and “new paradigm of digital transformation”. he pointed out that today's digitalization is equivalent to the electrification of the past, and alicloud is playing the role of providing centralized supply of arithmetic power. at present, alicloud's eighth-generation ecs instances based on cipu and fourth-generation intel xeon scalable processors have advantages such as high computing power, high io and high security, and can accelerate the industry with advanced computing power. and in a large number of digital transformation practices, alicloud also sees a path of self-evolution, which is summarized in the diamond model, i.e., taking business online, network collaboration, and decision-making intelligence as the key tasks, and continuously maintaining the spirit of innovation, agility, and openness to ultimately achieve the goal of data-driven development.

mr. wu yiming and mr. yang wei gave a presentation on “intel's innovative technology enabling the entire drug r&d process”. wu yiming said that more and more pharmaceutical companies are beginning to use ai to assist drug development. in order to drive ai applications to the ground, intel has recently released and will soon release a number of new products, in order to form a more complete general computing special acceleration product technology cornerstone. for example, when releasing the fourth-generation xeon scalable processors this year, it built in a variety of workload-specific gas pedals and introduced the xeon max series with integrated hbm (high bandwidth memory) to better address the arithmetic needs of memory bandwidth-sensitive applications. the upcoming habana gaudi 2 , is more suitable for large model applications. in addition, intel will further layout and evolution of gpu and fpga products to enrich the layout of heterogeneous computing power. at the same time, intel will also provide and update a variety of mainstream frameworks and tools optimized for intel architectures, so that developers can use them to rapidly develop and optimize a wide range of ai applications on intel hardware.

yang wei mainly introduced the typical optimization use cases of intel technology in the field of ai medicine. he said that in large molecule drug design, intel optimized alphafold2 last year based on the third-generation xeon scalable processor and achieved excellent results, and this year, after the launch of the xeon max series, intel made more optimization attempts based on this new product, and alphafold2 achieved more significant throughput improvement. in terms of small molecule drug design, molecule generation based on intel architecture enhancement can significantly improve inference throughput and save time and cost. in terms of large models, based on the ai acceleration capability of the fourth-generation intel xeon scalable processor, int4 can be used to support the inference of large models with parameters of about 700 million, which has a better performance in knowledge retrieval and drug counseling in the pharmaceutical field.

zhaorong li delivered a keynote speech on “exploration and practice of ai biological language large model” from three aspects: data, model and computing power. he pointed out that the success of alphafold is the inevitable result of data accumulation, and alicloud and intel optimized and tested alphafold2 based on the fourth-generation intel xeon scalable processor, and achieved good results in three scenarios: high-throughput, single-instance, and long sequence. in terms of modeling, alicloud annopro is now online. in addition, based on the alicloud sangshin application platform and intel's extensive technology portfolio, they have also optimized common applications such as gatk 4.0, relion 3.0, vasp and namd all on a cpu-based basis, achieving significant performance improvements in these applications.

at the conference, several other users and industry representatives from the pharmaceutical industry, academia, research and capital delivered speeches and held two roundtable discussions on hot topics such as the current status and prospects of ai medicine, aigc, big models, and investment considerations for ai medicine. they agreed that ai medicine is promising and can already play an important role in some fields, such as protein structure prediction, but overall, it needs to be further developed and matured in both technology and application.

the rapid evolution and application of advanced technologies such as big data, ai, cloud computing and 5g have driven profound changes in the pharmaceutical industry. china is now vigorously promoting the accelerated development from generic drugs to innovative drugs, which requires the head enterprises in the upstream and downstream of the pharmaceutical industry to join hands with experts in cloud computing, big data and ai technologies for cross-border cooperation and innovation. this conference is a positive response to this trend and demand, aiming to add bricks and mortar to the realization of “healthy china” and “digital china”.