Volcano Engine uses "four-wheel drive" to help "smart medicine future"

From June 22 to June 23, the third China Pharmaceutical Think Tank Forum with the theme of “Smart Medicine for the Future” was successfully held in Boao, Hainan.

As the organizer of the conference, Volcano Engine gathered with more than 100 experts from pharmaceutical companies, Internet companies, universities, laboratories, and research institutes to jointly look forward to the new trends in the development and transformation of artificial intelligence in the pharmaceutical industry, share new practical experiences in the implementation of cutting-edge technologies, and in-depth explore new ideas for artificial intelligence to accelerate the digital development of the pharmaceutical industry.

Using "four-wheel drive" to boost the development of "smart drugs"

In his speech titled "The "Four-Wheel Drive" of Computing Power, Data, Algorithms, and Knowledge to Accelerate Scientific Research Efficiency and Knowledge Discovery", Zhang Xin, Vice President of Volcano Engine, analyzed the current opportunities and challenges in the pharmaceutical industry, and explained the solutions and advantages of Volcano Engine's "Four-Wheel Drive" at different stages, injecting new AI power into accelerating R&D and improving scientific efficiency in the scientific research field.

Zhang Xin, Vice President of Volcano Engine

With the continuous increase in data volume, the scientific research field has entered a new paradigm of data-intensive and AI-accelerated R&D, and the emergence of large models has also brought multi-dimensional changes to the use, development and R&D methods of the pharmaceutical industry's technology field, as well as the efficiency experience. Zhang Xin said that in response to the new needs of pharmaceutical research in the AI ​​era, Volcano Engine has proposed a new "four-wheel drive" solution that includes computing power, data, algorithms, and knowledge, which can be generally divided into three stages for implementation.

In the first phase, Volcano Engine uses the biomedical operating system Bio-OS and Workspace as the core to integrate data, workflows, notebooks, execution engines and other content. Based on the technical advantages of cloud-native scheduling and support for multi-cloud heterogeneity, it makes data and computing traceable, shareable, collaborative and inheritable.

The second stage is the scientific research literature knowledge Copilot. Enterprises can improve the overall efficiency of scientific research institutions through literature and knowledge mining, and combine the explicit associations of knowledge to enhance the ability of large models to reason and discover knowledge, and quickly extract, interpret and create knowledge from massive information.

Entering the third stage, it is necessary to build a scientific research intelligent body Bio-Agent. Enterprises can use the language big model to realize automated experimental process orchestration and data output, and use the knowledge-driven model in the medical field to build a biomedical research agent to free scientific research innovation from constraints.

Big model, four steps to implement in the medical field

Yu Honglei, senior product director of Volcano Engine, pointed out in his keynote speech titled "Generative AI Helps Accelerate Innovation in Medicine and Health" that generative AI has formed a new industry map, and language models, multimodal models, and industry vertical models have set off a new wave of AI native applications, indicating that the turning point for big models to move from technology to application has arrived.

Yu Honglei, Senior Product Director of Volcano Engine

For pharmaceutical companies, the era of big models where "application is king" will also be a strategic opportunity to achieve transcendence. Volcano Engine has proposed a "four-step" approach to the implementation of big models in vertical fields by accumulating best practices, helping companies seize the initiative, quickly build applications in the pharmaceutical and health industry, and establish competitive advantages. Yu Honglei shared the specific implementation steps:

The first step is to choose the right base model. Based on the Doubao model family, Volcano Engine helps pharmaceutical companies build a strong base model capability, providing support in the popularization of medical knowledge, answering health questions, and diagnosing and treating diseases.

The second step is to accelerate the industrialization of models. Volcano Engine provides a GPU computing power pool of tens of thousands of GPUs, a full-stack AI development engineering optimization solution, and AI Infra stability management capabilities to ensure that enterprises can achieve efficient model reasoning and fine-tuning with computing power support with high utilization, high cost performance, high development experience, and low operation and maintenance deployment;

The third step is to lower the threshold for scenario-based application development. HiAgent, an AI application innovation platform launched by Volcano Engine, evolves the development method of large model applications from code to natural language description, allowing every employee in the enterprise to easily build intelligent entities and accelerate business innovation;

The fourth step is to achieve the data flywheel effect. The emergence of big models reduces the difficulty of data consumption and helps enterprises better form a two-way positive cycle from data assets to business applications. The faster the data flywheel rotates, the richer the scene and user cognition data obtained, the higher the accuracy of big model application, and the feedback of big model capabilities is realized.

Application scenarios of large models in the pharmaceutical industry are emerging rapidly and accelerating implementation

Nowadays, the application cases of big models are spreading rapidly. Taking the extraction of clinical program information as an example, Volcano Engine is using its big model capabilities to help a pharmaceutical research and development company (CRO) efficiently and high-quality build a professional knowledge base in the disease dimension to better serve clinical experts. At the same time, through workflow arrangement, key information of clinical trial programs is extracted and summarized to form informed consent forms for subjects. This process has been reduced from the original 6 hours of manual time to a few minutes now.

Intelligent assistants for medical enterprises are also an important application scenario of the big model. Hospitals can use AI robots to provide 7x24 general practice health consultation and intelligent medical guidance services, filter and match patient needs, and more accurately connect with doctors for registration. For chronic disease patients who have the right to choose their own medicines, medication consultation and guidance services are also provided.

In the field of medical content generation, AI text generation capabilities are used to help improve the efficiency of medical professional content production. Medical popular science content can be produced in large quantities for C-end users every day, while taking on the "propositional composition" of upstream tasks and conducting academic promotion for specific groups of people. The medical content generation service provided by Volcano Engine has covered most specialties, with an average daily PV of one million people.

As a cloud service platform under ByteDance, Volcano Engine is committed to becoming a powerful partner for customers' digital transformation and intelligent upgrades. In the future, Volcano Engine will continue to explore and analyze the application scenarios of large models, help medical research innovation, and help medical institutions upgrade to intelligence.

As a winner of Toutiao's Qingyun Plan and Baijiahao's Bai+ Plan, the 2019 Baidu Digital Author of the Year, the Baijiahao's Most Popular Author in the Technology Field, the 2019 Sogou Technology and Culture Author, and the 2021 Baijiahao Quarterly Influential Creator, he has won many awards, including the 2013 Sohu Best Industry Media Person, the 2015 China New Media Entrepreneurship Competition Beijing Third Place, the 2015 Guangmang Experience Award, the 2015 China New Media Entrepreneurship Competition Finals Third Place, and the 2018 Baidu Dynamic Annual Powerful Celebrity.